1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
6948
6949
6950
6951
6952
6953
6954
6955
6956
6957
6958
6959
6960
6961
6962
6963
6964
6965
6966
6967
6968
6969
6970
6971
6972
6973
6974
6975
6976
6977
6978
6979
6980
6981
6982
6983
6984
6985
6986
6987
6988
6989
6990
6991
6992
6993
6994
6995
6996
6997
6998
6999
7000
7001
7002
7003
7004
7005
7006
7007
7008
7009
7010
7011
7012
7013
7014
7015
7016
7017
7018
7019
7020
7021
7022
7023
7024
7025
7026
7027
7028
7029
7030
7031
7032
7033
7034
7035
7036
7037
7038
7039
7040
7041
7042
7043
7044
7045
7046
7047
7048
7049
7050
7051
7052
7053
7054
7055
7056
7057
7058
7059
7060
7061
7062
7063
7064
7065
7066
7067
7068
7069
7070
7071
7072
7073
7074
7075
7076
7077
7078
7079
7080
7081
7082
7083
7084
7085
7086
7087
7088
7089
7090
7091
7092
7093
7094
7095
7096
7097
7098
7099
7100
7101
7102
7103
7104
7105
7106
7107
7108
7109
7110
7111
7112
7113
7114
7115
7116
7117
7118
7119
7120
7121
7122
7123
7124
7125
7126
7127
7128
7129
7130
7131
7132
7133
7134
7135
7136
7137
7138
7139
7140
7141
7142
7143
7144
7145
7146
7147
7148
7149
7150
7151
7152
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7167
7168
7169
7170
7171
7172
7173
7174
7175
7176
7177
7178
7179
7180
7181
7182
7183
7184
7185
7186
7187
7188
7189
7190
7191
7192
7193
7194
7195
7196
7197
7198
7199
7200
7201
7202
7203
7204
7205
7206
7207
7208
7209
7210
7211
7212
7213
7214
7215
7216
7217
7218
7219
7220
7221
7222
7223
7224
7225
7226
7227
7228
7229
7230
7231
7232
7233
7234
7235
7236
7237
7238
7239
7240
7241
7242
7243
7244
7245
7246
7247
7248
7249
7250
7251
7252
7253
7254
7255
7256
7257
7258
7259
7260
7261
7262
7263
7264
7265
7266
7267
7268
7269
7270
7271
7272
7273
7274
7275
7276
7277
7278
7279
7280
7281
7282
7283
7284
7285
7286
7287
7288
7289
7290
7291
7292
7293
7294
7295
7296
7297
7298
7299
7300
7301
7302
7303
7304
7305
7306
7307
7308
7309
7310
7311
7312
7313
7314
7315
7316
7317
7318
7319
7320
7321
7322
7323
7324
7325
7326
7327
7328
7329
7330
7331
7332
7333
7334
7335
7336
7337
7338
7339
7340
7341
7342
7343
7344
7345
7346
7347
7348
7349
7350
7351
7352
7353
7354
7355
7356
7357
7358
7359
7360
7361
7362
7363
7364
7365
7366
7367
7368
7369
7370
7371
7372
7373
7374
7375
7376
7377
7378
7379
7380
7381
7382
7383
7384
7385
7386
7387
7388
7389
7390
7391
7392
7393
7394
7395
7396
7397
7398
7399
7400
7401
7402
7403
7404
7405
7406
7407
7408
7409
7410
7411
7412
7413
7414
7415
7416
7417
7418
7419
7420
7421
7422
7423
7424
7425
7426
7427
7428
7429
7430
7431
7432
7433
7434
7435
7436
7437
7438
7439
7440
7441
7442
7443
7444
7445
7446
7447
7448
7449
7450
7451
7452
7453
7454
7455
7456
7457
7458
7459
7460
7461
7462
7463
7464
7465
7466
7467
7468
7469
7470
7471
7472
7473
7474
7475
7476
7477
7478
7479
7480
7481
7482
7483
7484
7485
7486
7487
7488
7489
7490
7491
7492
7493
7494
7495
7496
7497
7498
7499
7500
7501
7502
7503
7504
7505
7506
7507
7508
7509
7510
7511
7512
7513
7514
7515
7516
7517
7518
7519
7520
7521
7522
7523
7524
7525
7526
7527
7528
7529
7530
7531
7532
7533
7534
7535
7536
7537
7538
7539
7540
7541
7542
7543
7544
7545
7546
7547
7548
7549
7550
7551
7552
7553
7554
7555
7556
7557
7558
7559
7560
7561
7562
7563
7564
7565
7566
7567
7568
7569
7570
7571
7572
7573
7574
7575
7576
7577
7578
7579
7580
7581
7582
7583
7584
7585
7586
7587
7588
7589
7590
7591
7592
7593
7594
7595
7596
7597
7598
7599
7600
7601
7602
7603
7604
7605
7606
7607
7608
7609
7610
7611
7612
7613
7614
7615
7616
7617
7618
7619
7620
7621
7622
7623
7624
7625
7626
7627
7628
7629
7630
7631
7632
7633
7634
7635
7636
7637
7638
7639
7640
7641
7642
7643
7644
7645
7646
7647
7648
7649
7650
7651
7652
7653
7654
7655
7656
7657
7658
7659
7660
7661
7662
7663
7664
7665
7666
7667
7668
7669
7670
7671
7672
7673
7674
7675
7676
7677
7678
7679
7680
7681
7682
7683
7684
7685
7686
7687
7688
7689
7690
7691
7692
7693
7694
7695
7696
7697
7698
7699
7700
7701
7702
7703
7704
7705
7706
7707
7708
7709
7710
7711
7712
7713
7714
7715
7716
7717
7718
7719
7720
7721
7722
7723
7724
7725
7726
7727
7728
7729
7730
7731
7732
7733
7734
7735
7736
7737
7738
7739
7740
7741
7742
7743
7744
7745
7746
7747
7748
7749
7750
7751
7752
7753
7754
7755
7756
7757
7758
7759
7760
7761
7762
7763
7764
7765
7766
7767
7768
7769
7770
7771
7772
7773
7774
7775
7776
7777
7778
7779
7780
7781
7782
7783
7784
7785
7786
7787
7788
7789
7790
7791
7792
7793
7794
7795
7796
7797
7798
7799
7800
7801
7802
7803
7804
7805
7806
7807
7808
7809
7810
7811
7812
7813
7814
7815
7816
7817
7818
7819
7820
7821
7822
7823
7824
7825
7826
7827
7828
7829
7830
7831
7832
7833
7834
7835
7836
7837
7838
7839
7840
7841
7842
7843
7844
7845
7846
7847
7848
7849
7850
7851
7852
7853
7854
7855
7856
7857
7858
7859
7860
7861
7862
7863
7864
7865
7866
7867
7868
7869
7870
7871
7872
7873
7874
7875
7876
7877
7878
7879
7880
7881
7882
7883
7884
7885
7886
7887
7888
7889
7890
7891
7892
7893
7894
7895
7896
7897
7898
7899
7900
7901
7902
7903
7904
7905
7906
7907
7908
7909
7910
7911
7912
7913
7914
7915
7916
7917
7918
7919
7920
7921
7922
7923
7924
7925
7926
7927
7928
7929
7930
7931
7932
7933
7934
7935
7936
7937
7938
7939
7940
7941
7942
7943
7944
7945
7946
7947
7948
7949
7950
7951
7952
7953
7954
7955
7956
7957
7958
7959
7960
7961
7962
7963
7964
7965
7966
7967
7968
7969
7970
7971
7972
7973
7974
7975
7976
7977
7978
7979
7980
7981
7982
7983
7984
7985
7986
7987
7988
7989
7990
7991
7992
7993
7994
7995
7996
7997
7998
7999
8000
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
8012
8013
8014
8015
8016
8017
8018
8019
8020
8021
8022
8023
8024
8025
8026
8027
8028
8029
8030
8031
8032
8033
8034
8035
8036
8037
8038
8039
8040
8041
|
<!DOCTYPE html><html><head><title>R: R News</title>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/katex@0.15.3/dist/katex.min.css">
<script>
const macros = { "\\R": "\\textsf{R}", "\\mbox": "\\text", "\\code": "\\texttt"};
function processMathHTML() {
var l = document.getElementsByClassName('reqn');
for (let e of l) { katex.render(e.textContent, e, { throwOnError: false, macros }); }
return;
}</script>
<script defer src="https://cdn.jsdelivr.net/npm/katex@0.15.3/dist/katex.min.js"
onload="processMathHTML();"></script>
<link rel="stylesheet" type="text/css" href="R.css">
</head><body><div class="container"><main>
<table style="width: 100%;"><tr><td>NEWS</td><td style="text-align: right;">R Documentation</td></tr></table>
<h2>R News</h2>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.5.3</h3>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>tools/fetch-recommended</code> can be used instead of
<code>tools/rsync-recommended</code> to fetch recommended packages
into <span class="rlang"><b>R</b></span> sources using <code>curl</code> on systems without
<code>rsync</code> or behind firewalls.
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> C++ standard specifications (<code>CXX_STD =</code> in
‘<span class="file">src/Makevars*</span>’ and in the <code>SystemRequirements</code> field of
the ‘<span class="file">DESCRIPTION</span>’ file) are now checked more thoroughly.
Invalid values are still ignored but now give a warning, as do
contradictory specifications.
</p>
</li>
<li><p> (Preliminary) support for C++26 has been extended to Windows.
</p>
</li>
<li><p> A non-zero exit status from ‘<span class="file">cleanup</span>’, ‘<span class="file">cleanup.win</span>’
or ‘<span class="file">cleanup.ucrt</span>’, if requested via options <span class="option">--clean</span> or
<span class="option">--preclean</span>, is now reported with a warning.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>all.equal(obj, simple, check.class=FALSE)</code> now is true, also
when <code>simple</code> is a bare atomic vector and <code>obj</code> has a simple
class, fixing the first part of <a href="https://bugs.R-project.org/show_bug.cgi?id=18971">PR#18971</a> thanks to Jan Gorecki.
</p>
</li>
<li> <p><code>str(x, give.attr=FALSE)</code> no longer shows attributes when
<code>x</code> is a zero length <code>"Date"</code> or <code>"POSIXt"</code> object.
</p>
</li>
<li><p> Tweaks to binning for <code>bw.SJ()</code> and <code>bw.ucv()</code> in the
very rare case of data which have an extremely small range compared to
their absolute values, e.g., <code>x <- 6e9 + 6:9</code>.
</p>
</li>
<li><p> Formatting book-type <code>bibentry</code> objects now converts
LaTeX accents also in the publisher and series fields.
</p>
</li>
<li> <p><code>model.frame(~1, list(), na.action=na.pass)</code> and similar
“border-line” uses no longer produce invalid data frames,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18977">PR#18977</a>, reported with patch by Mikael Jagan.
</p>
</li>
<li> <p><code>length(<POSIXlt>) <- v</code> is more careful about balancing,
notably when <code>v</code> is not integer, thanks to Suharto Anggono's
remarks on the mailing list R-devel.
</p>
</li>
<li> <p><code>approx(<x_with_ties>, <y_with_NA>, na.rm = FALSE)</code> now
should always call the <code>ties()</code> function, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17604">PR#17604</a>
reported by Bill Dunlap.
</p>
</li>
<li> <p><code>besselJ(1, 1e-15)</code> and similar now give correct results,
thanks to Leo Mada and other “R-help”ers.
</p>
</li>
<li> <p><code>vignette(<var>pkg</var>::<var>topic</var>)</code> is now a documented
usage variant and confines vignette retrieval to the specified package.
</p>
</li>
<li> <p><code>pretty(ch)</code> again correctly works with <code>ch</code> a
character vector of numbers.
</p>
</li>
<li> <p><code>persp()</code> labels the three axes correctly also when C level
<code>atan2pi()</code> is available, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=19007">PR#19007</a> by Klaus Schliep.
</p>
</li>
<li><p> Large (tall) subscripts in plotmath expressions are now
positioned correctly. They were being positioned too high.
</p>
</li>
<li> <p><code>Ops</code> group (Arith, Compare,..) methods for
matrix-<code>"ts"</code> now do return (zero length) matrices even when the
two series do not overlap, a very partial fix of <a href="https://bugs.R-project.org/show_bug.cgi?id=18972">PR#18972</a> which is to
be fully addressed later.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.5.2</h3>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> now handles archives with extension
‘<span class="file">.tar</span>’ or ‘<span class="file">.tar.zstd</span>’ (where <code>zstd</code> compression
is supported by the <span class="rlang"><b>R</b></span> build).
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>t.test(c(1:3, Inf))</code> and similar no longer produce an error but
return a (still not so useful) <code>"htest"</code> result, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18901">PR#18901</a>, thanks to Jesse Alderliesten.
</p>
</li>
<li> <p><code>attr(., "tsp") <- val</code> now uses <code>getOption("ts.eps")</code>
instead of hardwired <code>1e-5</code>; consequently, <code>ts(.., ts.eps=*)</code>
now passes <code>ts.eps</code> to the <code>"tsp"</code> setting C code;
both fixing a long-standing ‘FIXME’.
</p>
</li>
<li> <p><code>insertSource()</code> now ignores the internal
<code>.packageName</code> object, avoiding a superfluous message.
</p>
</li>
<li><p> In static HTML help, links to vignette files from the default
‘<span class="file">doc/index.html</span>’ page now also work for packages not installed
in the default <code>.Library</code>,
thanks to a report by Patrice Kiener.
</p>
</li>
<li>
<p><code>fixInNamespace("<var>S3method</var>")</code> failed to update the S3
methods table when the generic was not on the search path.
</p>
</li>
<li><p> In plain-text help (<code>Rd2txt</code>), an initial newline from
an Rd inline <code style="white-space: pre;">⁠\eqn⁠</code> no longer breaks the paragraph.
</p>
</li>
<li> <p><code>hist(*, log = "x")</code> now works without a warning, thanks
to Martin Smith's <a href="https://bugs.R-project.org/show_bug.cgi?id=18921">PR#18921</a>.
</p>
</li>
<li> <p>Subassigning <code>"POSIXlt"</code>, i.e., <code><tdat>[i] <- val</code>
and <code><tdat>[[i]] <- val</code> now rebalance as they should,
thanks to Mikael Jagan's <a href="https://bugs.R-project.org/show_bug.cgi?id=18919">PR#18919</a>.
</p>
</li>
<li> <p><code><POSIXlt>[*]</code> (re-)setting <code>"balanced"</code>, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18681">PR#18681</a> comment #7, thanks to Mikael Jagan.
</p>
</li>
<li><p> All four of <code>{col,row}{Sums,Means}(Z, na.rm=TRUE)</code> now correctly work
with complex <code>Z</code> where <code>is.na(Re(Z))</code> differs from <code>is.na(Im(Z))</code>,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18942">PR#18942</a>, unearthed by Dirk Eddelbuettel.
</p>
</li>
<li><p> Fix for glyph rendering on the <code>quartz()</code> device when
there is other (“normal”) text drawn on the device. The
problem was that the text transformation matrix was not reset
so glyphs would be rendered incorrectly (often completely outside
the device, i.e., not visible).
</p>
</li>
<li><p> Functions <code>install.packages()</code> and <code>download.packages()</code> again
consult option <code>download.file.method</code> when the download method is
unspecified.
</p>
</li>
<li> <p>Tanguy Barthelemy and colleagues at the
‘R Dev Day’ following Rencontres R in May 2025
extended the help page of <code>lm()</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18058">PR#18058</a>.
As suggested by Thomas Soeiro, such notes were also added to the
<code>glm()</code>, <code>poly()</code> and <code>splines::bs()</code> and <code>ns()</code>
pages.
</p>
</li>
<li> <p><code>lbeta(1i, 1)</code> now signals an error, as <code>lbeta()</code> is not implemented
for <code>complex</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18946">PR#18946</a> from Ben Bolker and Kasper Kristensen.
</p>
</li>
<li><p> The Cairo-based SVG device uses <code>pt</code> as the default
document unit also with Cairo >= 1.17.8 (<a href="https://bugs.R-project.org/show_bug.cgi?id=18912">PR#18912</a>).
</p>
</li>
<li> <p><code>pretty(*, eps.correct = 2)</code> has been fixed, e.g., to
avoid over 1 million length result for <code>pretty(c(0, 1e-322), eps.correct = 2)</code>.
</p>
</li>
<li><p> When <code>"POSIXlt"</code> date-time objects are <code>NA</code>-padded
from subsetting or increasing <code>length</code> in the <code>`[`</code> or
<code>`length<-`</code> methods, the <code>"balanced"</code> attribute is set
to <code>NA</code>, now; noted in <a href="https://bugs.R-project.org/show_bug.cgi?id=18681">PR#18681</a>, thanks to Mikael Jagan.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.5.1</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> The internal method of <code>unzip()</code> now follows <code>unzip 6.00</code>
in how it handles extracted file paths which contain <code>"../"</code>.
With thanks to Ivan Krylov.
</p>
</li></ul>
<h4>INSTALLATION</h4>
<ul>
<li><p> Standalone <code>nmath</code> can be built with early-2025
versions of <code>clang</code>-based compilers such as LLVM
<code>clang</code> 20, Apple <code>clang</code> 17 and Intel
<code>icx</code> 2025.0.
</p>
</li>
<li><p> Tcl/Tk 9 can be used to build package <code>tcltk</code>:
this has become the default in some Linux distributions.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Java detection in <code>javareconf</code> could not detect
‘<span class="file">libjvm.*</span>’ in the <code>zero</code> variant of the <abbr>JDK</abbr> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18884">PR#18884</a>).
All valid variants as of <abbr>JDK</abbr> 24u are now supported.
</p>
</li>
<li> <p><code>factanal(.., rotation=*)</code> now correctly updates
<code>rotmat</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18886">PR#18886</a>.
</p>
</li>
<li> <p><code>dnbinom(<large>, <muchlarger>, ..)</code> now is <code>0</code>
correctly, instead of <code>NaN</code> or <code>Inf</code> sometimes.
</p>
</li>
<li> <p><code>dbinom(<large>, n=Inf, ..)</code> is <code>0</code> now correctly,
instead of <code>NaN</code> which also fixes many <code>dnbinom()</code> cases,
notably those mentioned in <a href="https://bugs.R-project.org/show_bug.cgi?id=16727">PR#16727</a> comment #5.
</p>
</li>
<li><p> Fixing C level “binomial deviance” <code>bd0()</code> for
extreme arguments (preventing under-/overflow) solves more <a href="https://bugs.R-project.org/show_bug.cgi?id=16727">PR#16727</a>
cases and also prevents some full accuracy loss in such cases for
<code>dbinom()</code>, <code>dnbinom()</code>, and via <code>dbinom_raw()</code>
potentially <code>dgeom()</code>, <code>dhyper()</code>, <code>dbeta()</code>, and
<code>df()</code>.
</p>
</li>
<li> <p><code>signif(1.**e308, digits)</code> no longer truncates unnecessarily
(but still to prevent overflow to <code>Inf</code>), fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18889">PR#18889</a>.
</p>
</li>
<li> <p><code>prettyNum(*, zero.print={>=1-char}, replace.zero=TRUE)</code>
now works as documented, thanks to Marttila Mikko and
Ivan Krylov's messages on R-devel.
</p>
</li>
<li> <p><code>pbeta(x, a,b, ..)</code> for very large <code>a,b</code> no longer returns
<code>NaN</code> but the correct values (0 or 1, or their logs for <code>log.p = TRUE</code>).
This improves Mathlib's C level <code>bratio()</code> and hence also
<code>pnbinom()</code>, etc..
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.5.0</h3>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>as.integer(rl)</code> and hence <code>as.raw(rl)</code> now work for
a <code>list</code> of <code>raw(1)</code> elements, as proposed by Michael
Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18696">PR#18696</a>.
</p>
</li>
<li> <p><span class="pkg">graphics</span>' <code>grid()</code> gains optional argument
<code>nintLog</code>.
</p>
</li>
<li><p> New functions <code>check_package_urls()</code> and
<code>check_package_dois()</code> in package <span class="pkg">tools</span> for checking
URLs and <abbr>DOI</abbr>s in package sources.
</p>
</li>
<li><p> New <code>head()</code> and <code>tail()</code> methods for class
<code>"ts"</code> time series, proposed by Spencer Graves on R-devel.
</p>
</li>
<li><p> New <code>qr.influence()</code> function, a “bare bones”
interface to the <code>lm.influence()</code> leave-one-out diagnostics
computations; wished for in <a href="https://bugs.R-project.org/show_bug.cgi?id=18739">PR#18739</a>.
</p>
</li>
<li><p> Package <code>citation()</code> results auto-generated from the
package metadata now also provide package <abbr>DOI</abbr>s for CRAN and
Bioconductor packages.
</p>
</li>
<li><p> New function <code>grepv()</code> identical to <code>grep()</code> except
for the default <code>value = TRUE</code>.
</p>
</li>
<li> <p><code>methods(<pkg>:::<genfun>)</code> now does report methods when
neither the generic nor the methods have been exported.
</p>
</li>
<li> <p><code>pdf()</code> gains an <code>author</code> argument to set the
corresponding metadata field, and logical arguments
<code>timestamp</code> and <code>producer</code> to optionally omit the
respective metadata.
(Thanks to Edzer Pebesma.)
</p>
</li>
<li> <p><code>grDevices::glyphInfo()</code> gains a <code>rot</code> argument
to allow per-glyph rotation.
(Thanks to Daniel Sabanes Bove.)
</p>
</li>
<li><p> Package <span class="pkg">tools</span> now exports functions
<code>CRAN_current_db()</code>, <code>CRAN_aliases_db()</code>,
<code>CRAN_rdxrefs_db()</code>, <code>CRAN_archive_db()</code>, and
<code>CRAN_authors_db()</code>.
</p>
</li>
<li><p> Package <span class="pkg">tools</span> now exports functions
<code>R()</code> and <code>parse_URI_reference()</code>.
</p>
</li>
<li><p> Package <span class="pkg">tools</span> now exports functions
<code>base_aliases_db()</code> and <code>base_rdxrefs_db()</code>.
</p>
</li>
<li><p> It is now possible to set the background color for row and
column names in the data editor on Windows (<code>Rgui</code>).
</p>
</li>
<li> <p><code>Rterm</code> on Windows now accepts input lines of unlimited
length.
</p>
</li>
<li> <p><code>file.info()</code> on Windows now provides file owner name and
domain.
</p>
</li>
<li> <p><code>Sys.info()</code> on Windows now provides current user domain.
</p>
</li>
<li> <p><code>findInterval()</code> gets new arguments <code>checkSorted</code>
and <code>checkNA</code> which allow skipping relatively costly checks;
related to <a href="https://bugs.R-project.org/show_bug.cgi?id=16567">PR#16567</a>.
</p>
</li>
<li> <p><code>pnorm(x)</code> underflows more gracefully.
</p>
</li>
<li> <p><code>get(nam, env)</code> now signals a <em>classed</em> error,
<code>"getMissingError"</code>, as “subclass” of
<code>"missingArgError"</code> where the latter is used also in similar
situations, e.g., <code>f <- function(x) exp(x); try(f())</code> .
</p>
</li>
<li><p> The set operations now avoid the <code>as.vector()</code>
transformation for same-kind apparently vector-like operands.
</p>
</li>
<li> <p><code>md5sum()</code> can be used to compute an MD5 hash of a raw
vector of bytes by using the <code>bytes=</code> argument instead
of <code>files=</code>. The two arguments are mutually exclusive.
</p>
</li>
<li><p> Added function <code>sha256sum()</code> in package <span class="pkg">tools</span>
analogous to <code>md5sum()</code> implementing the <abbr>SHA</abbr>-256 hashing
algorithm.
</p>
</li>
<li><p> The <code>xtfrm()</code> method for class <code>"AsIs"</code> is now
considerably faster thanks to a patch provided by Ivan Krylov.
</p>
</li>
<li><p> The <code>merge()</code> method for data frames will no longer
convert row names used for indexing using <code>I()</code>, which will
lead to faster execution in cases where <code>sort = TRUE</code> and
<code>all.x</code> and/or <code>all.y</code> are set to <code>TRUE</code>.
</p>
</li>
<li><p> The <span class="pkg">methods</span> package internal function
<code>.requirePackage()</code> now calls <code>requireNamespace(p)</code>
instead of <code>require(p)</code>, hence no longer adding packages to
the <code>search()</code> path in cases methods or class definitions are
needed. Consequently, previous workflows relying on the old
behaviour will have to be amended by adding corresponding
<code>library(p)</code> calls.
</p>
</li>
<li><p> More <span class="rlang"><b>R</b></span>-level messages use a common format containing
<code>"character string"</code> for more consistency and less
translation work.
</p>
</li>
<li> <p><code>available.packages()</code> and <code>install.packages()</code> get
an optional switch <code>cache_user_dir</code>, somewhat experimentally.
</p>
</li>
<li><p> The <code>sunspot.month</code> data have been updated to Oct 2024;
because of recalibration also historical numbers are changed, and we
keep the previous data as <code>sunspot.m2014</code> for
reproducibility.
</p>
</li>
<li><p> The <code>quartz()</code> device now supports alpha masks. Thanks
to George Stagg, Gwynn Gebeyhu, Heather Turner, and
Tomek Gieorgijewski.
</p>
</li>
<li><p> The <code>print()</code> method for date-time objects (<code>POSIX.t</code>)
gets an optional <code>digits</code> argument for <em>fractional</em>
seconds, passed to improved <code>format.POSIXlt()</code>; consequently,
<code>print(<date.time>, digits = n)</code> allows to print fractions of
seconds.
</p>
</li>
<li> <p><code>install.packages()</code> and <code>download.packages()</code>
download packages simultaneously using <code>libcurl</code>, significantly
reducing download times when installing or downloading multiple
packages.
</p>
</li>
<li><p> Status reporting in <code>download.file()</code> has been extended to
report the outcome for individual files in simultaneous downloads.
</p>
</li>
<li><p> The Rd <code style="white-space: pre;">⁠\link⁠</code> macro now allows markup in the link text
when the topic is given by the optional argument, e.g.,
‘<span class="samp">⁠\link[=gamma]{\eqn{\Gamma(x)}}⁠</span>’.
</p>
</li>
<li><p> If <code>La_library()</code> is empty, <code>sessionInfo()</code> still
reports <code>La_version()</code> when available.
</p>
</li>
<li> <p><code>seq.int(from, to, by, ....)</code> when <code class="reqn">|by| = 1</code> now
behaves as if <code>by</code> was omitted, and hence returns <code>from:to</code>,
possibly integer.
</p>
</li>
<li> <p><code>seq.Date(from, to, by, ....)</code> and <code>seq.POSIXt(..)</code>
now also work when <code>from</code> is missing and sufficient further
arguments are provided, thanks to Michael Chirico's report, patch
proposal in <a href="https://bugs.R-project.org/show_bug.cgi?id=17672">PR#17672</a> and ‘R Dev Day’ contributions.
</p>
<p>The <code>Date</code> method also works for <code>seq(from, to)</code>, when
<code>by</code> is missing and now defaults to <code>"1 days"</code>.
</p>
<p>It is now documented (and tested) that the <code>by</code> string may be
<em>abbreviated</em> in both <code>seq</code> methods.
</p>
<p>Both methods return or keep internal type <code>"integer"</code> more
consistently now. Also, <code>as.POSIXct({})</code> is internally integer.
</p>
</li>
<li> <p><code>duplicated()</code>, <code>unique()</code>, and <code>anyDuplicated()</code>
now also work for class <code>expression</code> vectors.
</p>
</li>
<li><p> New function <code>use()</code> to use packages in R scripts with
full control over what gets added to the search path. (Actually
already available since <span class="rlang"><b>R</b></span> 4.4.0.)
</p>
</li>
<li><p> New connection type <code>zstdfile</code> for files compressed by
<code>zstd</code> if <span class="rlang"><b>R</b></span> was built with such support. <code>file()</code>
and <code>gzfile()</code> can automagically read such files.
</p>
</li>
<li> <p><code>memCompress()</code> and <code>memDecompress()</code> have options
to use <code>zstd</code> compression if <span class="rlang"><b>R</b></span> was built with support
for it.
</p>
</li>
<li><p> There is some support for <code>zstd</code> compression of
tarballs in <code>tar()</code> and <code>untar()</code>. (This depends on OS
support of <code>libzstd</code> or by <code>tar</code>.)
</p>
</li>
<li> <p><code>print(summary(<numbers>))</code> gets new optional argument
<code>zdigits</code> to allow more flexible and consistent (double)
rounding. The current default <code>zdigits = 4L</code> is somewhat
experimental. Specifying both <code>digits = *, zdigits = *</code> allows
behaviour independent of the global <code>digits</code> option.
</p>
</li>
<li><p> The <code>format()</code> method for <code>"difftime"</code> objects gets a
new back compatible option <code>with.units</code>.
</p>
</li>
<li><p> A <code>summary()</code> method for <code>"difftime"</code> objects which
prints nicely, similar to those for <code>"Date"</code> and <code>"POSIXct"</code>.
</p>
</li>
<li> <p><code>unique()</code>'s default method now also deals with
<code>"difftime"</code> objects.
</p>
</li>
<li> <p><code>optimize(f, *)</code> when <code>f(x)</code> is not finite says
more about the value in its <code>warning</code> message. It no longer
replaces <code>-Inf</code> by the largest <em>positive</em> finite number.
</p>
</li>
<li><p> The documentation of <code>gamma()</code> and <code>is.numeric()</code>
is more specific, thanks to the contributors of <a href="https://bugs.R-project.org/show_bug.cgi?id=18677">PR#18677</a>.
</p>
</li>
<li><p> New dataset <code>gait</code> thanks to Heather Turner and
Ella Kaye, used in examples.
</p>
</li>
<li><p> New datasets <code>penguins</code> and <code>penguins_raw</code> thanks to
Ella Kaye, Heather Turner, and Kristen Gorman.
</p>
</li>
<li> <p><code>isSymmetric(<matrix>)</code> gains a new option <code>trans = "C"</code>;
when set to non-default, it tests for “simple” symmetry of
complex matrices.
</p>
</li>
<li> <p><code>model.frame()</code> produces more informative error messages
in some cases when variables in the formula are not found, thanks to
Ben Bolker's <a href="https://bugs.R-project.org/show_bug.cgi?id=18860">PR#18860</a>.
</p>
</li>
<li> <p><code>selectMethod(f, ..)</code> now keeps the function name if the
function belongs to a group generic and the method is for the
generic.
</p>
</li></ul>
<h4>BLAS and LAPACK</h4>
<ul>
<li><p> The bundled BLAS and LAPACK sources have been updated to
those shipped as part of January 2025's LAPACK 3.12.1.
</p>
</li>
<li><p> It is intended that this will be the last update to BLAS and
LAPACK in the <span class="rlang"><b>R</b></span> sources. Those building <span class="rlang"><b>R</b></span> from source are
encouraged to use external BLAS and LAPACK and this will be required in
future.
</p>
</li>
<li><p> This update was mainly bug fixes but contained a barely
documented major change. The set of BLAS routines had been
unchanged since 1988, so throughout <span class="rlang"><b>R</b></span>'s history. This update
introduced new BLAS routines <code>dgemmtr</code> and <code>zgemmtr</code>
which are now used by LAPACK routines. This means that BLAS
implementations are no longer interchangeable.
</p>
</li>
<li><p> To work around this, <span class="rlang"><b>R</b></span> can be configured with option
<span class="option">--with-2025blas</span> which arranges for the 2025 BLAS
additions to be compiled into <code>libRlapack</code> (the internal
LAPACK, not built if an external LAPACK is used).
</p>
<p>This option allows the continuation of the practice of swapping
the BLAS in use by symlinking ‘<span class="file">lib/libRblas.*</span>’. It has
the disadvantage of using the reference BLAS version of the 2025
routines whereas an enhanced BLAS might have an optimized version
(OpenBLAS does as from version 0.3.29).
</p>
</li>
<li><p> Windows builds currently use the internal LAPACK and by
default the internal BLAS: notes on how to swap the latter
<em>via</em> ‘<span class="file">Rblas.dll</span>’ are in file
‘<span class="file">src/extra/blas/Makefile.win</span>’.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> A C23 compiler (if available) is now selected by default for
compilation of <span class="rlang"><b>R</b></span> and packages. <span class="rlang"><b>R</b></span> builds can opt out <em>via</em> the
<code>configure</code> flag <span class="option">--without-C23</span>, unless the
specified or default (usually <code>gcc</code>) compiler defaults to C23:
<code>gcc</code> 15 does.
</p>
<p>A C23 compiler is known to be selected with
<code>gcc</code> 13–15,
LLVM <code>clang</code> 18–20 (and 15 should),
Apple <code>clang</code> 15–17 and
Intel 2024.2–2025.0 (and 2022.2 should).
</p>
<p>Current binary distributions on macOS use Apple
<code>clang</code> 14 and so do not use C23.
</p>
</li>
<li><p> The minimum <code>autoconf</code> requirement for a maintainer
build has been increased to <code>autoconf</code> 2.72.
</p>
</li>
<li><p> Building the HTML and Info versions of the R manuals now
requires <code>texi2any</code> from Texinfo 6.1 or later.
</p>
</li>
<li><p> Failures in building the manuals under ‘<span class="file">doc</span>’ now abort
the installation, removing any file which caused the failure.
</p>
</li>
<li><p> Control of symbol visibility is now supported on macOS (the
previous check only worked on ELF platforms).
</p>
</li>
<li><p> There is now support for installing the debug symbols for
recommended packages on macOS: see ‘<span class="samp">⁠REC_INSTALL_OPT⁠</span>’ in file
‘<span class="file">config.site</span>’.
</p>
</li>
<li> <p><code>configure</code> is now able to find an external
<code>libintl</code> on macOS (the code from an older GNU gettext
distribution failed to try linking with the macOS Core Foundation
framework).
</p>
</li></ul>
<h4>INSTALLATION on WINDOWS</h4>
<ul>
<li><p> Both building <span class="rlang"><b>R</b></span> and installing packages use the C compiler
in C23 mode.
</p>
</li>
<li><p> R on Windows by default uses <code>pkg-config</code> for linking against
external libraries. This makes it easier to test R and packages with
alternative toolchains (such as from Msys2, e.g., testing with LLVM
and possibly with sanitizers). It also allows more significant Rtools
updates within a single R minor release.
</p>
</li>
<li><p> The installer scripts for Windows have been tailored to
Rtools45, an update of the Rtools44 toolchain. It is
based on GCC 14. The experimental support for 64-bit ARM
(<code>aarch64</code>) CPUs is based on LLVM 19. R-devel and <span class="rlang"><b>R</b></span>
4.5.x are no longer maintained to be buildable using
Rtools44 and it is advised to switch to Rtools45.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li> <p><code>is.R()</code> is defunct. Environment variable
<span class="env">_R_DEPRECATED_IS_R_</span> no longer has any effect.
</p>
</li>
<li><p> Deprecated (for more than 9 years!) functions
<code>linearizeMlist</code>, <code>listFromMlist</code>, and <code>showMlist</code> and
the <code>"MethodsList"</code> class for S4 method handling were removed
from package <span class="pkg">methods</span>.
Deprecated functions <code>balanceMethodsList</code>, <code>emptyMethodsList</code>,
<code>inheritedSubMethodLists</code>, <code>insertMethod</code>,
<code>insertMethodInEmptyList</code>, <code>makeMethodsList</code>,
<code>mergeMethods</code>, <code>MethodsList</code>, <code>MethodsListSelect</code>, and
<code>SignatureMethod</code> were made defunct, as were the
<code>"MethodsList"</code> branches of functions <code>assignMethodsMetaData</code>,
<code>finalDefaultMethod</code>, and <code>MethodAddCoerce</code>.
</p>
</li>
<li> <p><code>getMethods(*, table = TRUE)</code> is deprecated.
</p>
</li>
<li><p> Building with the bundled (and old) version of
<code>libintl</code> is deprecated and now gives a <code>configure</code>
warning. This should be selected only if neither the OS's
<code>libc</code> (as on GNU Linux) nor an external <code>libintl</code>
library provide suitable functions.
</p>
<p>Instead install <code>libintl</code> from a recent version of GNU
gettext (available for macOS) or use <code>configure</code> option
<span class="option">--disable-nls</span>.
</p>
<p>The ability to use the bundled version may be removed as soon as
<span class="rlang"><b>R</b></span> 4.5.1.
</p>
</li>
<li><p> The deprecated <code>xfig()</code> graphics device has been removed.
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> Packages are now installed using C23 where supported by the
OS and <span class="rlang"><b>R</b></span> build.
</p>
<p>Packages using <span class="rlang"><b>R</b></span>'s compiler settings can ask <strong>not</strong> to use C23
<em>via</em> including <code>USE_C17</code> in <code>SystemRequirements</code>
or can be installed by <code>R CMD INSTALL --use-C17</code>.
(Some packages ignore these settings in their <code>configure</code>
script or when compiling in sub-directories of ‘<span class="file">src</span>’, as
will those using a ‘<span class="file">src/Makefile</span>’.)
</p>
</li>
<li><p> Source installs now report the package version in the log.
</p>
</li>
<li><p> There is preliminary support for C++26 with GCC <code class="reqn">>=</code> 14, Apple
<code>clang++</code> <code class="reqn">>=</code> 16 and LLVM <code>clang++</code> <code class="reqn">>=</code> 17.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The non-API and hidden entry points <code>Rf_setIVector</code>,
<code>Rf_setRVector</code> and <code>Rf_setSVector</code> have been removed.
</p>
</li>
<li><p> The internal code for changing the parent of an environment
now signals an error if the new parent is not an environment or if
the change would create a cycle in the parent chain.
</p>
</li>
<li> <p><code>SET_TYPEOF</code> now signals an error unless the old and
new types have compatible memory structure and content. Use of
<code>SET_TYPE</code> in package C code should be avoided and may be
deprecated in the near future. It is better to allocate an object
of the desired type in the first place.
</p>
</li>
<li><p> The set of LAPACK (double and complex) routines declared in
the headers ‘<span class="file">R_ext/Lapack.h</span>’ and ‘<span class="file">R_ext/Applic.h</span>’ has
been extended, mostly to routines actually in use by packages.
</p>
</li>
<li><p> Memory allocation messages now use the (non-SI notation)
<code>"Mb"</code>, <code>"Gb"</code> , ..., and <code>"Mbytes"</code> strings as
<em>arguments</em> instead of as part of the (translatable format)
string. This is one step for <a href="https://bugs.R-project.org/show_bug.cgi?id=18297">PR#18297</a>; from Henrik Bengtsson.
</p>
</li>
<li><p> Header ‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’) now
always includes header ‘<span class="file">float.h</span>’ or ‘<span class="file">cfloat</span>’ for
constants such as <code>DBL_MAX</code>.
</p>
</li>
<li><p> Strict R headers are now the default. This removes the
legacy definitions of <code>PI</code>, <code>Calloc</code>, <code>Realloc</code> and
<code>Free</code>: use <code>M_PI</code>, <code>R_Calloc</code>, <code>R_Realloc</code> or
<code>R_Free</code> instead.
</p>
</li>
<li><p> The deprecated and seemingly never-used S-compatibility
macros <code>F77_COM</code> and <code>F77_COMDECL</code> have been removed
from header ‘<span class="file">R_ext/RS.h</span>’.
</p>
</li>
<li><p> The <code>enum</code> <code>Rboolean</code> defined in header
‘<span class="file">R_ext/Boolean.h</span>’ now has a fixed underlying type of
<code>int</code> on platforms whose C compiler supports this.
</p>
<p>This is a C23 feature (taken from C++11) and also supported in all
C standards by some versions of <code>clang</code> (from LLVM and
Apple) and (with a warning when using <span class="option">-pedantic</span>) by GCC
when in C17 mode.
</p>
<p>A fair amount of code has assumed this: it may be changed to a
smaller type in future. In particular, as standard compilers do
not check the validity of assignment to an <code>enum</code>, it has been
possible to assign <code>NA_INTEGER</code> to an <code>Rboolean</code>
variable, coerce it to <code>int</code> and recover the value.
</p>
<p>If there were a platform which used an underlying type of a
different size this would be an <abbr>ABI</abbr>-breaking change (but we are
unaware of any such platform).
</p>
</li>
<li><p> Header ‘<span class="file">R_ext/Boolean.h</span>’ now ensures that a <code>bool</code>
type is available either as a keyword (C23 and C++) or by
including the C99 header ‘<span class="file">stdbool.h</span>’. This is being used
internally in <span class="rlang"><b>R</b></span> to replace <code>Rboolean</code> by <code>bool</code>.
</p>
</li>
<li><p> There are new functions <code>asRboolean</code> and
<code>asBool</code>, variants of <code>asLogical</code> more suited to
converting logical arguments to <code>Rboolean</code> or to
<code>bool</code>. They require a length-one input and throw an error if
that evaluates to <code>NA</code>.
</p>
</li>
<li><p> Header ‘<span class="file">R_exts/Error.h</span>’ now ensures that
<code>Rf_error</code> and similar are given a <code>noreturn</code>
attribute when used from C++ under all compilers.
</p>
</li>
<li><p> Header ‘<span class="file">R_exts/Utils.h</span>’ no longer contains a
declaration for <code>F77_SUB(interv)</code>. This is intended to be
called from Fortran and was wrongly declared: <code>LOGICAL</code> in
Fortran corresponds to <code>int *</code> not <code>Rboolean *</code>.
</p>
</li>
<li><p> Defining <code>R_INCLUDE_BOOLEAN_H</code> to <code>0</code> before
including headers ‘<span class="file">R.h</span>’ or ‘<span class="file">Rinternals.h</span>’ (or any other
header which includes ‘<span class="file">R_ext/Boolean.h</span>’) stops the inclusion
of header <code>R_ext/Boolean.h</code> which ‘defines’ constants
<code>TRUE</code>, <code>FALSE</code>, <code>true</code>, <code>false</code> and the type
<code>bool</code> which some package maintainers wish to avoid.
</p>
<p>Note that the last three are keywords in C23 and C++11 so cannot be
avoided entirely. However, with commonly-used compilers they can be
masked by a macro of the same name, often with a warning.
</p>
</li></ul>
<h4>C-LEVEL API</h4>
<ul>
<li><p> The ‘Writing R Extensions’ Texinfo source now
contains very experimental annotations for more clearly
identifying the API status of C entry points. These annotations
are used to produce indices for API, experimental API, and
embedded API entry points in the rendered versions. This is very
preliminary and may be dropped if a better approach emerges.
</p>
<p>Also for Fortran-callable entry points which are part of the API.
</p>
</li>
<li> <p>‘Writing R Extensions’ has a new section
‘Moving into C API compliance’ to help package authors move
away from using non-API endpoints. This section will continue to
be updated as work on clarifying and tightening the C API
continues.
</p>
</li>
<li><p> New API function <code>R_mkClosure</code>. This checks that its
arguments are valid and should be used instead of
<code>allocSExp(CLOSXP</code> followed by <code>SET_FORMALS</code>,
<code>SET_BODY</code>, and <code>SET_CLOENV</code>.
</p>
</li>
<li><p> New API functions <code>R_ClosureFormals</code>,
<code>R_ClosureBody</code>, and <code>R_ClosureEnv</code> for extracting
closure components. The existing functions <code>R_ClosureExpr</code>
and <code>R_BytecodeExpr</code> have also been added to the API.
</p>
</li>
<li><p> New API function <code>R_ParentEnv</code> corresponding to <span class="rlang"><b>R</b></span>'s
<code>parent.env()</code>.
</p>
</li>
<li><p> Further non-API entry points have been added to those reported
by <code>R CMD check</code>: <code>COMPLEX0</code>, <code>ddfind</code>,
<code>DDVAL</code>, <code>ENSURE_NAMEDMAX</code>, <code>ENVFLAGS</code>,
<code>FRAME</code>, <code>HASHTAB</code>, <code>INTERNAL</code>, <code>IS_ASCII</code>,
<code>IS_UTF8</code>, <code>LEVELS</code>, <code>NAMED</code>, <code>PRSEEN</code>,
<code>RDEBUG</code>, <code>REAL0</code>, <code>Rf_findVarInFrame3</code>,
<code>SET_BODY</code>, <code>SET_CLOENV</code>, <code>SET_FORMALS</code>,
<code>SET_PRSEEN</code>, <code>SET_RDEBUG</code>, <code>STRING_PTR</code>,
<code>SYMVALUE</code>, and <code>VECTOR_PTR</code>. Any declarations for these
in public header files will be removed in the near future, and
they will be hidden where possible.
</p>
</li>
<li><p> Some <code>R CMD check</code> ‘<span class="samp">⁠NOTE⁠</span>’s on the use of
non-API entry points have been upgraded to ‘<span class="samp">⁠WARNING⁠</span>’s in
preparation for removing declarations and, where possible, hiding
these entry points.
</p>
</li>
<li><p> Additional non-API entry points have been added to those
reported by <code>R CMD check</code>: <code>IS_LONG_VEC</code>,
<code>PRCODE</code>, <code>PRENV</code>, <code>PRVALUE</code>, <code>R_nchar</code>,
<code>Rf_NonNullStringMatch</code>, <code>R_shallow_duplicate_attr</code>,
<code>Rf_StringBlank</code>, <code>SET_TYPEOF</code>, <code>TRUELENGTH</code>,
<code>XLENGTH_EX</code>, and <code>XTRUELENGTH</code>.
</p>
</li>
<li><p> Enable defining <span class="env">R_NO_REMAP_RMATH</span> and calling
<code>Rf_*()</code> as has been documented in
‘Writing R Extensions’ for a while, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18800">PR#18800</a>
thanks to Mikael Jagan and Suharto Anggono.
</p>
</li>
<li> <p><code>R_GetCurrentSrcref(skip)</code> now skips calls rather
than <code>srcref</code>s, consistent with counting items in the
<code>traceback()</code> display. If <code>skip == NA_INTEGER</code>,
it searches for the first <code>srcref</code>, starting at
the current evaluation state and proceeding through the
call stack; otherwise, it returns the <code>srcref</code>
of the requested entry from the call stack.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD INSTALL</code> (and hence <code>check</code>) now
compile C++ code with <code>-DR_NO_REMAP</code>.
</p>
<p>‘Writing R Extensions’ has been revised to describe the remapped
entry points, for with the <code>Rf_</code> prefix remains optional when
used from C code (but is recommended for new C code).
</p>
</li>
<li> <p><code>R CMD check --as-cran</code> notes bad parts in the
‘<span class="file">DESCRIPTION</span>’ file's URL fields.
</p>
</li>
<li> <p><code>R CMD check</code> now reports more warnings on
long-deprecated/obsolete Fortran features reported by
<code>gfortran -Wall</code>. For hints on how to modernize these,
see
<a href="https://fortranwiki.org/fortran/show/Modernizing+Old+Fortran">https://fortranwiki.org/fortran/show/Modernizing+Old+Fortran</a>.
</p>
</li>
<li><p> Since almost all supported <span class="rlang"><b>R</b></span> versions now use UTF-8,
<code>R CMD check</code> no longer by default reports on marked
UTF-8 or Latin-1 strings in character data. Set environment
variable <span class="env">_R_CHECK_PACKAGE_DATASETS_SUPPRESS_NOTES_</span> to a
false value for the previous behaviour.
</p>
</li>
<li> <p><code>tools::checkDocFiles()</code> notes more cases of
usage documentation without corresponding <code style="white-space: pre;">⁠\alias⁠</code>.
</p>
</li>
<li> <p><code>R CMD check</code> with a true value for environment variable
<span class="env">_R_CHECK_BASHISMS_</span> checks more thoroughly, including for
<code>bash</code> scripts and bashisms in components of
<code>autoconf</code>-generated <code>configure</code> scripts.
</p>
</li>
<li> <p><code>R CMD check</code> gains option <span class="option">--run-demo</span> to
check demo scripts analogously to tests. This includes a check
for undeclared package dependencies: it can also be enabled
separately by setting the environment variable
<span class="env">_R_CHECK_PACKAGES_USED_IN_DEMO_</span> to a true value
(as done by <code>R CMD check --as-cran</code>).
</p>
</li>
<li> <p><code>R CMD build</code> now supports <span class="option">--compression=zstd</span>
on platforms with sufficient support for <code>zstd</code>.
</p>
</li>
<li> <p><code>tools::texi2pdf(..., texinputs=<var>paths</var>)</code> now
<em>pre</em>pends the specified <var>paths</var> to <span class="env">TEXINPUTS</span>.
When building <span class="rlang"><b>R</b></span> itself (‘<span class="file">doc/NEWS.pdf</span>’ and base vignettes)
or package manuals using <code>R CMD Rd2pdf</code>, it is ensured
that this <span class="rlang"><b>R</b></span>'s ‘<span class="file">Rd.sty</span>’ takes precedence over any other
(incompatible) versions in default “texmf trees”.
</p>
</li>
<li> <p><code>tools::Rd2latex()</code> no longer outputs an
‘<span class="samp">⁠\inputencoding{utf8}⁠</span>’ line by default;
such a declaration is obsolete since LaTeX 2018-04-01.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>update_pkg_po()</code> now copies ‘<span class="file">.mo</span>’ files to the
<span class="pkg">translation</span> package even if a ‘<span class="file">DESCRIPTION</span>’ file
exists, thanks to Michael Chirico fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18694">PR#18694</a>.
</p>
</li>
<li><p> Auto-generated <code>citation()</code> entries no longer include
(additional) URLs in the ‘<span class="samp">⁠note⁠</span>’ field (<a href="https://bugs.R-project.org/show_bug.cgi?id=18547">PR#18547</a>).
</p>
</li>
<li> <p><code>as.data.frame.list()</code> gets a new option <code>new.names</code>
and now preserves <code>NA</code> names, thus fixing the <code>format()</code>
method for data frames, and also bug <a href="https://bugs.R-project.org/show_bug.cgi?id=18745">PR#18745</a>.
Relatedly, the <code>format()</code> method gets an option <code>cut.names</code>.
</p>
</li>
<li> <p><code>stem()</code> formats correctly also in cases where rounding
up, e.g., from 9.96 to 10 needs more digits; thanks to Ella
Kaye and Kelly Bodwin, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=8934">PR#8934</a> during ‘R
Dev Day’ at useR!2024.
</p>
<p>Additionally, <code>stem(x)</code> now works normally also when
<code>length(x) == 1</code>.
</p>
</li>
<li> <p><span class="pkg">tools</span>' <code>toTitleCase()</code> now works better, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18674">PR#18674</a>, thanks to Shannon Pileggi, Sarah Zeller,
Reiko Okamoto, and Hugo Gruson's ‘R Dev Day’
effort.
</p>
</li>
<li><p> Printing matrices (typically with many rows and or columns)
now also omits columns when desirable according to option
<code>max.print</code>, or argument <code>max</code>, respectively.
This is primarily the work of Lorena Abad,
Ekaterina Akimova, Hanne Oberman, Abhishek Ulayil,
and Lionel Henry at the ‘R Dev Day’, thus fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=15027">PR#15027</a>.
</p>
</li>
<li> <p><code>Sys.setLanguage()</code> now warns about <em>some</em>
failures to change the language.
</p>
</li>
<li><p> Printing <code>ls.str()</code> now shows <code>"<missing>"</code> even when
<span class="rlang"><b>R</b></span>'s language setting is not English.
</p>
</li>
<li> <p><code>xyTable()</code> now handles and reports <code>NA</code>s
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18654">PR#18654</a>.
Thanks to Heather Turner and Zhian Kamvar for report and patch.
</p>
</li>
<li> <p><code>as(*, "raw")</code> now works as documented, thanks to Mikael
Jagan's <a href="https://bugs.R-project.org/show_bug.cgi?id=18795">PR#18795</a>.
</p>
</li>
<li><p> Informational messages of e.g., <code>print(1:1e4, max=1000)</code>,
now correctly mention <code>max</code> in addition to
<code>getOption("max.print")</code>.
</p>
</li>
<li> <p><code>rowSums(A, dims = dd)</code>, <code>colMeans(..)</code>, etc, give a
more helpful error message when <code>dd</code> is not of length one,
thanks to Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18811">PR#18811</a>.
</p>
</li>
<li> <p><code>seq.Date()</code> no longer explicitly coerces results from
integer to double, analogously with <code>seq.default()</code>,
<code>seq.int()</code> and <code>seq.POSIXt()</code>, resolving a <em>modified</em>
<a href="https://bugs.R-project.org/show_bug.cgi?id=18782">PR#18782</a>.
</p>
</li>
<li> <p><code>axisTicks(usr, ...)</code> documentation clarification for
<code>log = TRUE</code>, fixing bug <a href="https://bugs.R-project.org/show_bug.cgi?id=18821">PR#18821</a> thanks to Duncan Murdoch.
</p>
</li>
<li> <p><code>debug()</code> and <code>debugonce(fun)</code> now also accept a
string <code>fun</code> when it names an S4 generic, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18822">PR#18822</a>
thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>debugonce(<S4-simple-body>, signature=*)</code> now works
correctly when “called twice”, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18824">PR#18824</a> thanks to
Mikael Jagan.
</p>
</li>
<li> <p><code>format(dtime, digits=* / format=*)</code> is more consistent
when the <code>POSIXt</code> date-time object <code>dtime</code> has fractional
(non integer) seconds. Fixes <a href="https://bugs.R-project.org/show_bug.cgi?id=17350">PR#17350</a>, thanks to new contributions
by LatinR's ‘R Dev Day’ participants, Heather
Turner and Dirk Eddelbuettel; also fixes more cases, notably
when <code>format</code> contains ‘<span class="samp">⁠%OS<nodigit>⁠</span>’.
</p>
</li>
<li> <p><code>options(scipen = NULL)</code> and other invalid values now
signal an error instead of invalidating ops relying on a finite
integer value. Values outside the range -9 .. 9999 are now warned
about and set to a boundary or to the default <code>0</code>, e.g., in
case of an <code>NA</code>. This also fixes <a href="https://bugs.R-project.org/show_bug.cgi?id=16322">PR#16322</a>.
</p>
</li>
<li> <p><code>cbind()</code> could segfault with <code>NULL</code> inputs.
(Seen when <span class="rlang"><b>R</b></span> was built with <code>gcc-14</code>, <abbr>LTO</abbr> and C99
inlining semantics.)
</p>
</li>
<li><p> Fix segfault on <code>quartz()</code> from <code>grid.glyph()</code>
call with <code>glyphInfo()</code> that describes non-existent font
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18784">PR#18784</a>). Thanks to Tomek Gieorgijewski.
</p>
</li>
<li> <p><code>format()</code> etc, using <code>decimal.mark = s</code>, by default
getting <code>s <- getOption("OutDec")</code>, signals a warning (to become
an error in the future) when <code>s</code> is not a string with exactly
one character.
</p>
</li>
<li><p> When <code>s <- getOption("OutDec")</code> is not a string of one
character, a warning is signalled now whenever it is used in internal
C code, notably when calling the default methods of <code>format()</code>.
</p>
</li>
<li> <p><code>pwilcox()</code> and <code>qwilcox()</code> now check for user
interrupt less frequently.
</p>
</li>
<li> <p><code>summary(<stl>)</code> (which prints directly) finally gets the
same <code>digits</code> default as the formatting printing of default
<code>summary()</code> method results, and it is documented explicitly.
</p>
</li>
<li> <p><code>options(show.error.locations = TRUE)</code> once
again shows the most recent known location when an
error is reported. Setting its value to <code>"bottom"</code>
is no longer supported. Numerical values are converted
to logical.
</p>
</li>
<li><p> C API function <code>R_GetCurrentSrcref(skip)</code> now
returns <code>srcref</code> entries correctly. (Note that there
is also a change to the interpretation of <code>skip</code>;
see the C-LEVEL API entry above.)
</p>
</li>
<li> <p><code>tools::Rd2latex()</code> now removes leading and trailing
spaces from <code style="white-space: pre;">⁠\alias⁠</code> entries as documented, fixing indexing
and linking hiccups in some PDF manuals.
</p>
</li>
<li> <p><code>R CMD Rd2pdf</code> can now render the package manual from
a <span class="option">--latex</span> installation also when the help contains figures.
</p>
</li>
<li><p> The argument of <code>as.environment()</code> is now named <code>x</code>,
not <code>object</code>, as was always documented and shown when
printing it; thanks to Gael Millot's <a href="https://bugs.R-project.org/show_bug.cgi?id=18849">PR#18849</a>.
</p>
</li>
<li><p> When <code>R CMD check</code> aims at getting the time+date from a
world clock, it is more robust against unexpected non-error results,
thanks to Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18852">PR#18852</a>.
</p>
</li>
<li><p> The <code>tools::parseLatex()</code> parser made several parsing
errors (<a href="https://bugs.R-project.org/show_bug.cgi?id=18855">PR#18855</a>).
</p>
</li>
<li><p> Error messages produced by <code>tools::parseLatex()</code> are
now more readable (<a href="https://bugs.R-project.org/show_bug.cgi?id=18855">PR#18855</a>).
</p>
</li>
<li> <p><code>R CMD build <pkg></code> excludes more file patterns
when it tars the <pkg> directory, fixing both <a href="https://bugs.R-project.org/show_bug.cgi?id=18432">PR#18432</a> and
<a href="https://bugs.R-project.org/show_bug.cgi?id=18434">PR#18434</a>, for <code>vim</code> and <code>GNU Global</code>
<code>emacs</code> users, thanks to Dirk Eddelbuettel's patch.
</p>
</li>
<li> <p><code>quantile()</code>'s default method gets an option <code>fuzz</code>
to protect against floating point inaccuracy before rounding, thus
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=15811">PR#15811</a> and, en passant, <a href="https://bugs.R-project.org/show_bug.cgi?id=17683">PR#17683</a>.
</p>
</li>
<li><p> Printing arrays now also omits columns, rows and slices when
desirable according to option <code>max.print</code>, or argument
<code>max</code>, respectively, addressing most of the remaining part of
<a href="https://bugs.R-project.org/show_bug.cgi?id=15027">PR#15027</a>, thanks to Sherry Zhang's patch.
</p>
</li>
<li> <p><code>drop.terms(y ~ w, 1)</code> and similar now work, thanks to
Benjamin Sommer's report in <a href="https://bugs.R-project.org/show_bug.cgi?id=18861">PR#18861</a> and collaboration with
Heather Turner improving <code>reformulate()</code>.
</p>
</li>
<li><p> Many arguments which should be length-1 logical are checked
more thoroughly. The most commonly seen errors are in
<code>unlink(, recursive)</code>, <code>tempdir()</code> and the <code>na.rm</code>
arguments of <code>max()</code>, <code>min()</code>, <code>sum()</code>, ....
</p>
<p><code>grep()</code>, <code>strsplit()</code> and similar took
non-<code>TRUE</code>/<code>FALSE</code>
values of their logical arguments as <code>FALSE</code>, but these were
almost always mismatches to unnamed arguments and are now reported
as <code>NA</code>.
</p>
</li>
<li> <p><code>vignette("reshape")</code> is now also available on Windows.
</p>
</li>
<li> <p><code>trace(coerce, ..)</code> now works correctly, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18823">PR#18823</a>
thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>R CMD check</code> now also reports bad symbols in
package shared objects linked in from local static libraries
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18789">PR#18789</a>).
</p>
</li>
<li> <p><code>factanal()</code> now works correctly also, e.g., for
<span class="pkg">GPArotation</span>, <code>oblimin()</code> rotations, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18417">PR#18417</a>,
thanks to Coen Bernaards and others.
</p>
</li>
<li><p> Setting <code>attributes</code> on primitive functions is
deprecated now and already an error in the development version of <span class="rlang"><b>R</b></span>.
Changing the <code>environment</code> of a primitive does no longer happen
and signals a warning.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.4.3</h3>
<h4>INSTALLATION</h4>
<ul>
<li> <p><span class="rlang"><b>R</b></span> can be installed using C23 (for example with
<span class="option">-std=gnu23</span> or <span class="option">-std=gnu2x</span>) with recent compilers
including <code>gcc</code> 12–14, Apple <code>clang</code> 15–16, LLVM
<code>clang</code> 17–20 and Intel <code>icx</code> 2024.2.
</p>
<p>It can be installed with the upcoming (at the time of writing)
<code>gcc</code> 15, which defaults to C23.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The functions <code>R_strtod</code> and <code>R_atof</code> now allow
hexadecimal constants without an exponent, for compatibility with
their C99 versions (<a href="https://bugs.R-project.org/show_bug.cgi?id=18805">PR#18805</a>).
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD build</code> and <code>R CMD check</code> now allow
reference output for demo scripts (‘<span class="file">demo/<var>demo</var>.Rout.save</span>’
files) to be shipped with the package, as proposed by
Torsten Hothorn in <a href="https://bugs.R-project.org/show_bug.cgi?id=18816">PR#18816</a>.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>kappa(A, exact=TRUE)</code> for singular <code>A</code> returns
<code>Inf</code> more generally, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18817">PR#18817</a> reported by
Mikael Jagan.
</p>
</li>
<li><p> Fixed URLs of the sun spots (<code>sunspot.month</code> etc) data
sets and mention future changes due to recalibration.
</p>
</li>
<li><p> The parser now accepts hexadecimal constants with a decimal
point without an exponent (taken as <code>p0</code>) as documented in
<code>?NumericConstants</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18819">PR#18819</a>).
</p>
</li>
<li> <p><code>rbind()</code> now works correctly when inputs include a raw
vector and a logical, integer or double vector – previously the
inclusion of the latter was garbled.
</p>
</li>
<li> <p><code>smooth.spline()</code> checks validity of its arguments
<code>df.offset</code> and <code>penalty</code>: it could segfault if they
were <code>NULL</code>.
</p>
</li>
<li> <p><code>isGeneric(<primitive>, fdef=*, getName=TRUE)</code> now also
returns the name instead of just <code>TRUE</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18829">PR#18829</a>
reported by Mikael Jagan.
</p>
</li>
<li> <p><code>isGeneric(fdef = print)</code> now works, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18369">PR#18369</a>
thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>sort(x, method = "qsort")</code> made illegal accesses when
<code>x</code> has length 0.
</p>
</li>
<li> <p><code>dir.create()</code> is protected against being passed an
empty string as its <code>path</code> argument.
</p>
</li>
<li><p> Silent integer overflow could occur in the
‘exact’ computations for <code>fisher.test()</code> for
unrealistic inputs: this is now an error.
</p>
</li>
<li><p> Some invalid C-level memory accesses are avoided for
<code>loglin(, margin = NULL)</code>.
</p>
<p><code>loglin(, param = TRUE)</code> no longer gives an error in corner
cases such as a one-dimensional input.
</p>
</li>
<li> <p><code>dev.capabilities() $ events</code> now reports <code>"Idle"</code> if
the device provides it, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18836">PR#18836</a>, thanks to Trevor Davis.
</p>
</li>
<li> <p><code>arima(.., seasonal = <wrong-vector>)</code> correctly errors
now, ditto for <code>arima0()</code>, thanks to Norbert Kuder's report
on the R-devel list.
</p>
</li>
<li> <p><code>binomial(<link>)$linkinv(eta)</code> and <code>.. $mu.eta(eta)</code>
now also work for <code>"logit"</code> link when <code>is.integer(eta)</code>.
</p>
</li>
<li> <p><code>as.roman(x)</code> now should work platform independently,
also for, e.g., <code>x = "IIIII"</code> (= V) and <code>x = "IIIIII"</code>
(= VI).
</p>
</li>
<li> <p><code>R CMD Rd2pdf</code> works again on an installed package
directory containing LaTeX help (from option <span class="option">--latex</span>),
thanks to a report by Peter Ruckdeschel.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.4.2</h3>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The S-compatibility macros <code>F77_COM</code> and
<code>F77_COMDECL</code> defined in header ‘<span class="file">R_ext/RS.h</span>’ are
deprecated and will be removed shortly. We could find no record
of their use.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p>Mathlib function <code>lgammacor(x)</code> no longer warns about
underflow to zero for large <code>x</code>.
</p>
</li>
<li><p> Text widths and heights were incorrectly reported by the
Quartz device if the drawing context didn't exist yet (typically
when drawing off-screen to a window that is yet to appear, see
<a href="https://bugs.R-project.org/show_bug.cgi?id=18591">PR#18591</a>).
</p>
</li>
<li><p> The Quartz device could segfault in cases where paths with
spaces are used in the new glyph drawing API.
Thanks to Tomek Gieorgijewski (<a href="https://bugs.R-project.org/show_bug.cgi?id=18758">PR#18758</a>).
</p>
</li>
<li><p> On macOS in R CRAN builds, it is again possible to read
little-endian UTF-16 text with a <abbr>BOM</abbr> from a connection using
<code>encoding="UTF-16"</code>. Users building R from source should avoid
using the system <code>libiconv</code> in macOS 14.1 and later.
</p>
</li>
<li> <p><span class="pkg">methods</span>' internal <code>.requirePackage()</code> now re-enables
primitive method dispatch when needed; thanks to Ivan Krylov
for demystifying CRAN package check failures on the R-devel mailing
list.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.4.1</h3>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> Functions <code>R_atof</code> and <code>R_strtod</code> declared in
header ‘<span class="file">R_ext/Utils.h</span>’ are now documented in
‘Writing R Extensions’ and so formally part of the API.
</p>
</li>
<li><p> The non-API entry points <code>Rf_setSVector</code>,
<code>Rf_StringFalse</code>, <code>Rf_StringTrue</code> and
<code>Rf_isBlankString</code> have been added to those reported by
<code>R CMD check</code>.
</p>
</li>
<li><p> The new function <code>Rf_allocLang</code> is now available. This
provides an alternative to the idiom of calling
<code>Rf_allocList</code> followed by <code>SET_TYPEOF</code>.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> now reports as warnings what
<code>gfortran</code> calls ‘Fortran 2018 deleted features’, all
of which have long been marked as ‘obsolescent’ and some of
which were deleted in Fortran 2008 or earlier. Fortran compilers
are no longer required to support these.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>as.numeric()</code>, <code>scan()</code>, <code>type.convert()</code>
and other places which use the internal C function <code>R_strtod</code> now
require a <em>non-empty</em> digit sequence in a decimal or binary
exponent. This aligns with the C/POSIX standard for <code>strtod</code> and
with <code>?NumericConstants</code>.
</p>
</li>
<li> <p><code>as.data.frame(m, make.names=NA)</code> now works correctly for
a matrix <code>m</code> with <code>NA</code>'s in row names.
</p>
</li>
<li><p> The error message from <code><POSIXlt>[["hour"]]</code> and similar
now mentions <code>*[[, "hour"]]</code>, as wished for in <a href="https://bugs.R-project.org/show_bug.cgi?id=17409">PR#17409</a> and
proposed by Michael Chirico.
</p>
</li>
<li> <p><code>qbinom()</code> and potentially <code>qpois()</code>, <code>qnbinom()</code>,
no longer sometimes fail accurate inversion (of <code>pbinom()</code>,
etc), thanks to Christopher Chang's report and patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18711">PR#18711</a>.
</p>
</li>
<li><p> The internal help server on Windows can again serve requests
sent in quick succession, fixing a regression in <span class="rlang"><b>R</b></span> 4.4.0.
</p>
</li>
<li> <p><code>debugcall(<var>S3Generic</var>())</code> now also works when a
corresponding S4-generic version is in the <span class="pkg">methods</span> cache
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18143">PR#18143</a>).
</p>
</li>
<li><p> Package <span class="pkg">tools</span>' <code>toTitleCase(ch0)</code> now returns
<code>character(0)</code> when <code>ch0</code> is of zero length; fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18724">PR#18724</a>, reported by David Hugh Jones.
</p>
</li>
<li> <p><code>R CMD check</code> is no longer broken (without a check
result and no explanation in ‘<span class="file">00check.log</span>’) for a package
which declares an invalid <code>VignetteBuilder</code> in
‘<span class="file">DESCRIPTION</span>’ but has no vignettes.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.4.0</h3>
<h4>SIGNIFICANT USER-VISIBLE CHANGES</h4>
<ul>
<li><p> Startup banners, <code>R --version</code>, <code>sessionInfo()</code>
and <code>R CMD check</code> no longer report <code>(64-bit)</code> as part of
the platform as this is almost universal – the increasingly rare
32-bit platforms will still report <code>(32-bit)</code>.
</p>
<p>On Windows, ditto for window titles.
</p>
</li>
<li> <p><code>is.atomic(NULL)</code> now returns <code>FALSE</code>, as <code>NULL</code>
is not an atomic vector. Strict back-compatibility would
replace <code>is.atomic(foo)</code> by <code>(is.null(foo) || is.atomic(foo))</code>
but should happen only sparingly.
</p>
</li></ul>
<h4>NEW FEATURES</h4>
<ul>
<li><p> The <code>confint()</code> methods for <code>"glm"</code> and
<code>"nls"</code> objects have been copied to the <span class="pkg">stats</span>
package. Previously, they were stubs which called versions in
package <span class="pkg">MASS</span>. The <span class="pkg">MASS</span> namespace is no longer loaded
if you invoke (say) <code>confint(glmfit)</code>. Further, the
<code>"glm"</code> method for <code>profile()</code> and the <code>plot()</code>
and <code>pairs()</code> methods for class <code>"profile"</code> have been
copied from <span class="pkg">MASS</span> to <span class="pkg">stats</span>. (<code>profile.nls()</code> and
<code>plot.profile.nls()</code> were already in <span class="pkg">stats</span>.)
</p>
</li>
<li><p> The <code>confint()</code> and <code>profile</code> methods for
<code>"glm"</code> objects have gained a possibility to do profiling
based on the Rao Score statistic in addition to the default
Likelihood Ratio. This is controlled by a new <code>test =</code>
argument.
</p>
</li>
<li><p> The <code>pairs()</code> method for <code>"profile"</code> objects has
been extended with a <code>which =</code> argument to allow plotting only
a subset of the parameters.
</p>
</li>
<li><p> The <code>"glm"</code> method for <code>anova()</code> computes test
statistics and p-values by default, using a chi-squared test or
an F test depending on whether the dispersion is fixed or
free. Test statistics can be suppressed by giving argument
<code>test</code> a false logical value.
</p>
</li>
<li><p> In <code>setRepositories()</code> the repositories can be set
using their names via <code>name =</code> instead of index <code>ind =</code>.
</p>
</li>
<li> <p><code>methods()</code> and <code>.S3methods()</code> gain a
<code>all.names</code> option for the (rare) case where functions starting
with a ‘<span class="samp">⁠.⁠</span>’ should be included.
</p>
</li>
<li><p> Serializations can now be interrupted (e.g., by <kbd>Ctrl-C</kbd> on
a Unix-alike) if they take too long, e.g., from <code>save.image()</code>,
thanks to suggestions by Ivan Krylov and others on R-devel.
</p>
</li>
<li><p> New startup option <span class="option">--max-connections</span> to set the
maximum number of simultaneous connections for the session.
Defaults to 128 as before: allowed values up to 4096 (but resource
limits may in practice restrict to smaller values).
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> on Windows (since Windows 10 2004) now uses the new Segment
Heap allocator. This may improve performance of some memory-intensive
applications.
</p>
</li>
<li><p> When <span class="rlang"><b>R</b></span> packages are built, typically by <code>R CMD build <pkg></code>,
the new <code>--user=<build_user></code> option overrides the
(internally determined) user name, currently <code>Sys.info()["user"]</code>
or <span class="env">LOGNAME</span>. This is a (modified) fulfillment of Will Landau's
suggestion in <a href="https://bugs.R-project.org/show_bug.cgi?id=17530">PR#17530</a>.
</p>
</li>
<li> <p><code>tools::testInstalledBasic()</code> gets new optional
arguments <code>outDir</code> and <code>testSrcdir</code>, e.g., allowing to use
it in a <code><var>builddir</var> != <var>srcdir</var></code> setup, and in standard
“binary” Windows installation <b>if</b> a source ‘<span class="file">tests/</span>’
folder is present.
</p>
</li>
<li> <p><code>range(<DT_with_Inf>, finite = TRUE)</code> now work for objects
of class <code>"Date"</code>, <code>"POSIXct"</code>, and <code>"POSIXlt"</code> with
infinite entries, analogously to <code>range.default()</code>, as proposed
by Davis Vaughan on R-devel. Other <code>range()</code>-methods can make
use of new function <code>.rangeNum()</code>.
</p>
</li>
<li><p> New <code>.internalGenerics</code> object complementing
<code>.S3PrimitiveGenerics</code>, for documentation and low-level
book-keeping.
</p>
</li>
<li> <p><code>grid()</code> now invisibly returns the x- and y- coordinates
at which the grid-lines were drawn.
</p>
</li>
<li> <p><code>norm(., type)</code> now also works for complex matrices.
</p>
</li>
<li> <p><code>kappa(., exact = TRUE, norm = *)</code> now works for all
norms and also for complex matrices. In symmetric / triangular
cases, the new argument <code>uplo = "U" | "L"</code> allows the
upper or lower triangular part to be specified.
</p>
</li>
<li> <p><code>memDecompress(type = "unknown")</code> recognizes compression
in the default ‘zlib’ format as used by
<code>memCompress(type = "gzip")</code>.
</p>
</li>
<li> <p><code>memCompress()</code> and <code>memDecompress()</code> will use the
<code>libdeflate</code> library
(<a href="https://github.com/ebiggers/libdeflate">https://github.com/ebiggers/libdeflate</a>) if installed. This
uses the same type of compression for <code>type = "gzip"</code> but is
1.5-2x faster than the system <code>libz</code> library on some
common platforms: the speed-up may depend on the library version.
</p>
</li>
<li> <p><code>diff()</code> for objects of class <code>"Date"</code>,
<code>"POSIXct"</code>, and <code>"POSIXlt"</code> accepts a <code>units</code>
argument passed via <code>...</code>.
</p>
</li>
<li><p> Dynamic help now does a much better job of rendering package
‘<span class="file">DESCRIPTION</span>’ metadata.
</p>
</li>
<li> <p><code>Rprof()</code> gains an <code>event</code> argument and support
for elapsed (real) time profiling on Unix (<a href="https://bugs.R-project.org/show_bug.cgi?id=18076">PR#18076</a>).
</p>
</li>
<li> <p><code>filled.contour()</code> gains a <code>key.border</code> argument.
</p>
</li>
<li> <p><code>tools::update_pkg_po()</code> gets arguments <code>pot_make</code>
and <code>mo_make</code> for <em>not</em> re-making the corresponding
files, and additionally a <code>verbose</code> argument.
</p>
</li>
<li><p> Hexadecimal string colour specifications are now accepted
in short form, so, for example, we can use <code>"#123"</code>, which is
equivalent to <code>"#112233"</code>.
</p>
<p>Thanks to MikeFC for the original idea and Ella Kaye,
Malcolm Barrett, George Stagg, and Hanne Oberman for
the patch.
</p>
</li>
<li><p> Plain-text help shows <code style="white-space: pre;">⁠\var⁠</code> markup by angle brackets.
</p>
</li>
<li><p> The new experimental primitive function <code>declare()</code> is
intended to eventually allow information about <span class="rlang"><b>R</b></span> code to be
communicated to the interpreter, compiler, and code analysis
tools. The syntax for declarations is still being developed.
</p>
</li>
<li><p> Functions <code>psmirnov()</code>, <code>qsmirnov()</code> and
<code>rsmirnov()</code> in package <span class="pkg">stats</span> have had argument
<code>two.sided</code> renamed to <code>alternative</code>, to take into
account that the permutation distributions of the one-sided
statistics can be different in the case of ties. Consequence of
<a href="https://bugs.R-project.org/show_bug.cgi?id=18582">PR#18582</a>.
</p>
</li>
<li> <p><code>sort()</code> is now an implicit S4 generic in <span class="pkg">methods</span>.
</p>
</li>
<li><p> Formatting and printing, <code>format(z), print(z)</code>, of
complex vectors <code>z</code> no longer zap relatively small real or
imaginary parts to zero, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=16752">PR#16752</a>. This is an API
change, as it was documented previously to round real and
imaginary parts together on purpose, producing nicer looking
output. As mentioned, e.g. in the PR, this change is
compatible with many other “<span class="rlang"><b>R</b></span>-like” programming
environments.
</p>
<p>We have simplified the internal code and now basically format the
real and imaginary parts independently of each other.
</p>
</li>
<li><p> New experimental functions <code>Tailcall()</code> and
<code>Exec()</code> to support writing stack-space-efficient recursive
functions.
</p>
</li>
<li><p> Where characters are attempted to be plotted by
<code>pdf()</code>, <code>postscript()</code> and <code>xfig()</code> which are
not in the selected 8-bit character set (most often Latin-1) and
the <span class="rlang"><b>R</b></span> session is using a UTF-8 locale, the warning messages will
show the UTF-8 character rather than its bytes and one dot will be
substituted per character rather than per byte. (Platforms whose
<code>iconv()</code> does transliteration silently plot the
transliteration.)
</p>
<p>In a UTF-8 locale some transliterations are now done with a
warning (e.g., dashes and Unicode minus to hyphen, ligatures are
expanded, permille (‘<span class="samp">⁠‰⁠</span>’) is replaced by ‘<span class="samp">⁠o/oo⁠</span>’),
although the OS may have got there first. These are warnings as
they will continue to be replaced by dots in earlier versions of
<span class="rlang"><b>R</b></span>.
</p>
</li>
<li><p> The matrix multiplication functions <code>crossprod()</code> and
<code>tcrossprod()</code> are now also primitive and S3 generic, as
<code>%*%</code> had become in <span class="rlang"><b>R</b></span> 4.3.0.
</p>
</li>
<li> <p><code>source()</code> and <code>example()</code> have a new optional
argument <code>catch.aborts</code> which allows continued evaluation of
the <span class="rlang"><b>R</b></span> code after an error.
</p>
</li>
<li><p> The non-Quartz <code>tiff()</code> devices allow additional types
of compression if supported by the platform's ‘<span class="samp">⁠libtiff⁠</span>’ library.
</p>
</li>
<li><p> The list of base and recommended package names is now provided by
<code>tools::standard_package_names()</code>.
</p>
</li>
<li> <p><code>cairo_pdf()</code> and <code>cairo_ps()</code> default to
<code>onefile = TRUE</code> to closer match <code>pdf()</code> and
<code>postscript()</code>.
</p>
</li>
<li><p> New option <code>catch.script.errors</code> provides a documented way
to catch errors and then continue in non-interactive use.
</p>
</li>
<li> <p><code>L %||% R</code> newly in <span class="pkg">base</span> is an expressive idiom
for the phrases <code>if(!is.null(L)) L else R</code> or
<code>if(is.null(L)) R else L</code>.
</p>
</li>
<li><p> The return value from <code>warnings()</code> now always inherits
from <code>"warnings"</code> as documented, now also in the case of no
warnings where it previously returned <code>NULL</code>.
</p>
</li>
<li> <p><code>as.complex("1i")</code> now returns <code>0 + 1i</code> instead of
<code>NA</code> with a warning.
</p>
</li>
<li> <p><code>z <- c(NA, 1i)</code> now keeps the imaginary part
<code>Im(z[1]) == 0</code>, no longer coercing to <code>NA_complex_</code>.
Similarly, <code>cumsum(z)</code> correctly sums real and imaginary parts
separately, i.e., without “crosstalk” in case of <code>NA</code>s.
</p>
</li>
<li><p> On Alpine Linux <code>iconv()</code> now maps <code>"latin2"</code>,
<code>"latin-2"</code>, <code>"latin9"</code> and <code>"latin-9"</code> to encoding
names the OS knows about (case-insensitively).
</p>
</li>
<li> <p><code>iconv(sub = "Unicode")</code> now always zero-pads to four
(hex) digits, rather than to 4 or 8. (This seems to have become
the convention once Unicode restricted the number of Unicode
points to <code class="reqn">2^{21} - 1</code> and so will never need more
than 6 digits.)
</p>
</li>
<li> <p><code>NCOL(NULL)</code> now returns 0 instead of 1, for
consistency with <code>cbind()</code>.
</p>
</li>
<li><p> The information for the Euro glyph missing from the Adobe
‘<span class="file">.afm</span>’ files for the Courier, Helvetica and Times families
has been copied from their URW equivalents – this will
improve vertical centring in the <code>pdf()</code> and
<code>postscript()</code> devices.
</p>
</li>
<li><p> The included BLAS sources have been updated to those shipped
with LAPACK version 3.12.0. The changes are almost entirely cosmetic.
</p>
</li>
<li><p> The included LAPACK sources have been updated to version
3.12.0 and some further double-complex routines added.
</p>
</li>
<li><p> There are new font families for the 2014–5 URW 2.0 fonts
(see <code>?pdf</code>) which are included in recent versions of
Ghostscript. These have font widths for most Greek glyphs and a
few others which were missing from the original versions (whose
font families remain available for reproducibility, although
Ghostscript-based viewers will render using the 2.0 versions).
</p>
</li>
<li><p> Improve the large-n efficiency of <code>as.matrix(<dist>)</code>,
thanks an R contributors effort, notably by Tim Taylor and
Heather Turner, see <a href="https://bugs.R-project.org/show_bug.cgi?id=18660">PR#18660</a>.
</p>
</li>
<li><p> The default and <code>numeric</code> methods of <code>all.equal()</code>
get a <code>check.class</code> option.
</p>
</li>
<li> <p><code>zapsmall()</code> gets new optional arguments, function
<code>mFUN</code> and <code>min.d</code>, for extra flexibility; fulfills a
wish in <a href="https://bugs.R-project.org/show_bug.cgi?id=18199">PR#18199</a>. Also, it is now an implicit S4 generic in
package <span class="pkg">methods</span>.
</p>
</li>
<li><p> The Rd filter for <code>aspell()</code> gains an <code>ignore</code>
argument.
</p>
</li>
<li><p> New generic function <code>sort_by()</code>, primarily useful for
the <code>data.frame</code> method which can be used to sort rows of a
data frame by one or more columns.
</p>
</li>
<li><p> The licence headers for the <abbr>RPC</abbr> code in
‘<span class="file">src/extra/xdr</span>’ have been updated to use the GPL-compatible
licence published by Oracle America in 2010.
</p>
</li>
<li><p> New function <code>pkg2HTML()</code> in <span class="pkg">tools</span> to create
single-page HTML reference manuals for R packages.
</p>
</li>
<li><p> The byte code evaluator now uses less C stack space for
recursive calls to byte-compiled functions. It also makes more of
an effort to avoid allocations for scalar return values.
</p>
</li>
<li><p> New completion option <code>backtick</code> (disabled by default)
allows non-syntactic completions to be wrapped in backquotes. This
is currently only useful for Jupyter notebooks via the
<span class="pkg">IRkernel</span> package, and may cause problems for other backends.
</p>
</li>
<li><p> The numeric version creators now stop on invalid
non-character version specifications.
</p>
</li></ul>
<h4>INSTALLATION</h4>
<ul>
<li><p> The parser has been updated to work with <code>bison</code> 3.8.2,
which is now used for the pre-generated parsers in ‘<span class="file">gram.c</span>’,
<code>file.c</code>, and ‘<span class="file">gramRd.c</span>’. A few parser error messages
have changed, which may affect code that relies on exact messages.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> System valgrind headers are now required to use
<code>configure</code> option <span class="option">--with-valgrind-instrumentation</span>
with value <code>1</code> or <code>2</code>.
</p>
</li>
<li> <p><code>configure</code> will warn if it encounters a 32-bit
build, as that is nowadays almost untested.
</p>
</li>
<li><p> Environment variable <span class="env">R_SYSTEM_ABI</span> is no longer used and
so no longer recorded in ‘<span class="file">etc/Renviron</span>’ (it was not on Windows
and was only ever used when preparing package <span class="pkg">tools</span>).
</p>
</li>
<li><p> If the <code>libdeflate</code> library and headers are available,
<code>libdeflate</code> rather than <code>libz</code> is used to (de)compress
<span class="rlang"><b>R</b></span> objects in lazy-load databases. Typically tasks spend up to 5% of
their time on such operations, although creating lazy-data
databases is one of the exceptions.
</p>
<p>This can be suppressed if the library is available by the
<code>configure</code> option
<span class="option">--without-libdeflate-compression</span>.
</p>
</li>
<li> <p><code>configure</code> option <span class="option">--enable-lto=check</span> has
not worked reliably since 2019 and has been removed.
</p>
</li>
<li><p> The minimum <code>autoconf</code> requirement for a maintainer
build has been increased to <code>autoconf</code> 2.71.
</p>
<p>It is intended to increase this to 2.72 for <span class="rlang"><b>R</b></span> 4.5.0: the
distributed ‘<span class="file">configure</span>’ file was generated using 2.72.
</p>
</li>
<li><p> The minimum version requirement for an external LAPACK has
been reduced to 3.9.0.
</p>
</li>
<li><p> No default C++ compiler is set if no C++17 compiler is
detected: there is no longer an automatic fallback to C++14 or
C++11.
</p>
<p>Compilers from the last five years should have sufficient
support: for others macros <code>CXX</code> and <code>CXXSTD</code> can be set
in file ‘<span class="file">config.site</span>’ to provide a fallback if needed.
</p>
<p>The Objective-C++ compiler now by default uses the standard
selected by <span class="rlang"><b>R</b></span> for C++ (currently C++17) rather than the default
standard for the C++ compiler (which on macOS is still C++98).
</p>
</li></ul>
<h4>INSTALLATION on macOS</h4>
<ul>
<li><p> A new <code>configure</code> option <span class="option">--with-newAccelerate</span>
makes use of Apple's ‘new’ BLAS / LAPACK interfaces in their
Accelerate framework. Those interfaces are only available in macOS
13.3 or later, and building requires <abbr>SDK</abbr> 13.3 or later (from the
Command Line Tools or Xcode 14.3 or later).
</p>
<p>By default the option uses new Accelerate for BLAS calls: to also
use it for LAPACK use option <span class="option">--with-newAccelerate=lapack</span>. The
later interfaces provide LAPACK 3.9.1 rather than 3.2.1: 3.9.1 is
from 2021-04 and does not include the improved algorithms
introduced in LAPACK 3.10.0 (including for BLAS calls).
</p>
</li></ul>
<h4>INSTALLATION on WINDOWS</h4>
<ul>
<li><p> The makefiles and installer scripts for Windows have been
tailored to Rtools44, an update of the Rtools43 toolchain. It
is based on GCC 13 and newer versions of MinGW-W64, binutils
and libraries (targeting 64-bit Intel CPUs). R-devel can no longer be
built using Rtools43 without changes.
</p>
</li>
<li> <p>Rtools44 has experimental support for 64-bit ARM
(<code>aarch64</code>) CPUs <em>via</em> the LLVM 17 toolchain using
<code>lld</code>, <code>clang</code>/<code>flang-new</code> and <code>libc++</code>.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> notes when S4-style exports are used
without declaring a strong dependence on package <span class="pkg">methods</span>.
</p>
</li>
<li> <p><code>tools::checkRd()</code> (used by <code>R CMD check</code>)
detects more problems with <code style="white-space: pre;">⁠\Sexpr⁠</code>-based dynamic content,
including bad nesting of <code style="white-space: pre;">⁠\Sexpr⁠</code>s and invalid arguments.
</p>
</li>
<li> <p><code>tools::checkRd()</code> now reports Rd titles and section
names ending in a period; this is ignored by <code>R CMD check</code>
unless environment variable <span class="env">_R_CHECK_RD_CHECKRD_MINLEVEL_</span> is
set to -5 or smaller.
</p>
</li>
<li> <p><code>R CMD check</code> now notes Rd files without an
<code style="white-space: pre;">⁠\alias⁠</code>, as long documented in ‘Writing R Extensions’
§1.3.1.
The check for a missing <code style="white-space: pre;">⁠\description⁠</code> has been moved from
<code>tools::checkRd()</code> to <code>tools::checkRdContents()</code>.
</p>
</li>
<li> <p><code>R CMD check</code> now visits ‘<span class="file">inst/NEWS.Rd</span>’
and OS-specific ‘<span class="file">man</span>’ subdirectories when checking Rd files.
</p>
</li>
<li> <p><code>tools::checkDocFiles()</code> and
<code>tools::checkRdContents()</code> now also check internal Rd files
by default, but “specially” (ignoring missing documentation
of arguments).
</p>
</li>
<li> <p><code>R CMD Rdiff</code> gets option <span class="option">--useEx</span>.
</p>
</li>
<li> <p><code>R CMD check</code> now warns on non-portable uses of
Fortran <code>KIND</code> such as <code>INTEGER(KIND=4)</code> and
<code>REAL(KIND=8)</code>.
</p>
<p>To see the failing lines set environment variable
<span class="env">_R_CHECK_FORTRAN_KIND_DETAILS_</span> to a true value.
</p>
</li>
<li><p> When checking Rd files, <code>R CMD check</code> now
notes some of the “lost braces” that
<code>tools::checkRd()</code> finds. Typical problems are Rd macros
missing the initial backslash (e.g., ‘<span class="samp">⁠code{...}⁠</span>’), in-text
set notation (e.g., ‘<span class="samp">⁠{1, 2}⁠</span>’, where the braces need
escaping), and <code style="white-space: pre;">⁠\itemize⁠</code> lists with <code style="white-space: pre;">⁠\describe⁠</code>-style
entries of the form <code style="white-space: pre;">⁠\item{label}{description}⁠</code>.
</p>
</li>
<li> <p><code>R CMD INSTALL</code> (and hence <code>check</code>) will
compile C++ code with <code>-DR_NO_REMAP</code> if environment variable
<span class="env">_R_CXX_USE_NO_REMAP_</span> is set to a true value. It is planned
that this will in future become the default for compiling C++.
</p>
</li>
<li><p> The new built-in Rd macro <code style="white-space: pre;">⁠\dontdiff{}⁠</code> can be used
to mark example code whose output should be ignored when comparing
check output to reference output
(‘<span class="file">tests/Examples/<var>pkg</var>-Ex.Rout.save</span>’).
The <code style="white-space: pre;">⁠\dontdiff⁠</code> tag, like <code style="white-space: pre;">⁠\donttest⁠</code>, is <em>not</em>
shown on the rendered help page, so provides a clean alternative
to ‘<span class="samp">⁠## IGNORE_RDIFF_(BEGIN|END)⁠</span>’ comments.
</p>
</li>
<li> <p><code>R CMD build</code> when there is no ‘<span class="file">NAMESPACE</span>’, now
uses the recommended <code>exportPattern("^[^.]")</code>, instead of
exporting everything.
</p>
</li>
<li> <p><code>R CMD check</code> now warns about non-ASCII characters
in the ‘<span class="file">NAMESPACE</span>’ file (in addition to <span class="rlang"><b>R</b></span> files). Such
packages are not portable and may fail to install on some platforms.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> Headers ‘<span class="file">R_ext/Applic.h</span>’ and ‘<span class="file">R-ext/Linpack.h</span>’
used to include ‘<span class="file">R_ext/BLAS.h</span>’ although this was undocumented and
unneeded by their documented entry points. They no longer do so.
</p>
</li>
<li><p> New function <code>R_missing()</code>, factored out from
<code>do_missing()</code>, used to fix <a href="https://bugs.R-project.org/show_bug.cgi?id=18579">PR#18579</a>.
</p>
</li>
<li> <p><code>SEXP</code> type <code>S4SXP</code> has been renamed to <code>OBJSXP</code>
to support experimenting with alternative object systems. The
<code>S4SXP</code> value can still be used in <code>C</code> code but is now
deprecated. Based on contributions from the R Consortium's
Object-Oriented Programming Working Group.
</p>
</li>
<li><p> New function <code>pow1p(x,y)</code> for accurate <code>(1+x)^y</code>.
</p>
</li>
<li> <p><code>mkCharLenCE</code> was incorrectly documented to take a
<code>size_t</code> length but was implemented with <code>int</code> (and
character strings in <span class="rlang"><b>R</b></span> are limited to <code class="reqn">2^{31} - 1</code> bytes).
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li> <p><code>data()</code> no longer handles zipped data from
long-defunct (since <span class="rlang"><b>R</b></span> 2.13.0) <span class="option">--use-zip-data</span>
installations.
</p>
</li>
<li><p> The legacy graphics devices <code>pictex()</code> and
<code>xfig()</code> are now deprecated. They do not support recent
graphics enhancements and their font handling is rudimentary. The
intention is to retain them for historical interest as long as
they remain somewhat functional.
</p>
</li>
<li><p> Support for <code>encoding = "MacRoman"</code> has been removed
from the <code>pdf()</code> and <code>postscript()</code> devices – this was
a legacy encoding supporting classic macOS up to 2001 and no
longer has universal <code>libiconv</code> support.
</p>
</li>
<li> <p><code>is.R()</code> is deprecated as no other S dialect is known
to be in use (and this could only identify historical dialects,
not future ones).
</p>
<p>Further information on calls can be obtained by setting the
environment variable <span class="env">_R_DEPRECATED_IS_R_</span> to ‘<span class="samp">⁠error⁠</span>’
which turns the deprecation warning into an error and so by
default gives a traceback. (This is done by <code>R CMD check
--as-cran</code>.)
</p>
</li>
<li> <p><code>UseMethod()</code> no longer forwards local variables
assigned in the generic function into method call environments
before evaluating the method body. This makes method calls behave
more like standard function calls and makes method code easier to
analyze correctly.
</p>
</li>
<li><p> The twelve <code>as.data.frame.<class>()</code> methods which were
deprecated only via <span class="env">_R_CHECK_AS_DATA_FRAME_EXPLICIT_METHOD_</span> and
in <code>R CMD check --as-cran</code> are formally deprecated now in
favour of calling <code>as.data.frame()</code> or <code>as.data.frame.vector()</code>.
The deprecation “check” now works also when <code>as.data.frame()</code>
is S4 generic thanks to Ivan Krylov.
</p>
</li>
<li><p> The default method for the directional comparison operators
<code><</code>, <code>></code>, <code><=</code>, and <code>>=</code> now signals an error
when one of the operands is a language object, i.e. a symbol or a
call.
</p>
</li>
<li><p> For <code>terms.formula()</code>, deprecate <code>abb</code> and <code>neg.out</code>
arguments <em>formally</em> in addition to just documenting it.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> The <span class="pkg">methods</span> package is more robust to not being attached to
the search path. More work needs to be done.
</p>
</li>
<li> <p><code>pairwise.t.test()</code> misbehaved when subgroups had 0 <abbr>DF</abbr>
for variance, even with <code>pool.sd = TRUE</code>. (<a href="https://bugs.R-project.org/show_bug.cgi?id=18594">PR#18594</a> by
Jack Berry.)
</p>
</li>
<li><p> Probability distribution functions <code>[dpq]<distrib>(x, *)</code>,
but also <code>bessel[IKJY](x, .)</code> now consistently preserve
<code>attributes(x)</code> when <code>length(x) == 0</code>, e.g., for a
<code class="reqn">2 \times 0</code> matrix, thanks to Karolis Koncevičius'
report <a href="https://bugs.R-project.org/show_bug.cgi?id=18509">PR#18509</a>.
</p>
</li>
<li><p> Group “Summary” computations such as
<code>sum(1:3, 4, na.rm = 5, NA, 7, na.rm = LL)</code> now give an error
instead of either <code>17</code> or <code>NN</code> for <code>LL</code> true or false,
as proposed by Ivan Krylov on the R-devel mailing list. (This
also means it is now an error to specify <code>na.rm</code> more than once.)
</p>
</li>
<li> <p><code>as.complex(x)</code> now returns <code>complex(real = x,
imaginary = 0)</code> for <em>all</em> numerical and logical <code>x</code>,
notably also for <code>NA</code> or <code>NA_integer_</code>.
</p>
</li>
<li><p> Directories are now omitted by
<code>file.copy(, recursive = FALSE)</code> and in <code>file.append()</code>
(<a href="https://bugs.R-project.org/show_bug.cgi?id=17337">PR#17337</a>).
</p>
</li>
<li> <p><code>gsub()</code> and <code>sub()</code> are now more robust to integer
overflow when reporting errors caused by too large input strings
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18346">PR#18346</a>).
</p>
</li>
<li><p> Top-level handlers are now more robust to attempts to remove a
handler whilst handlers are running (<a href="https://bugs.R-project.org/show_bug.cgi?id=18508">PR#18508</a>).
</p>
</li>
<li><p> The handling of <code>Alt+F4</code> in dialogs created on Windows using
GraphApp has been fixed (<a href="https://bugs.R-project.org/show_bug.cgi?id=13870">PR#13870</a>).
</p>
</li>
<li> <p><code>density()</code> more consistently computes grid values for the
FFT-based convolution, following Robert Schlicht's analysis and
proposal in <a href="https://bugs.R-project.org/show_bug.cgi?id=18337">PR#18337</a>, correcting density values typically by a
factor of about 0.999. Argument <code>old.coords = TRUE</code> provides
back compatibility.
</p>
</li>
<li> <p><code>palette.colors()</code> gains a <code>name</code> argument that
defaults to <code>FALSE</code> controlling whether the vector of colours
that is returned has names (where possible). <a href="https://bugs.R-project.org/show_bug.cgi?id=18529">PR#18529</a>.
</p>
</li>
<li> <p><code>tools::xgettext()</code> no longer extracts the
(non-translatable) class names from <code>warningCondition</code> and
<code>errorCondition</code> calls.
</p>
</li>
<li> <p><code>S3method(<gen>, <class>, <func>)</code> in the ‘<span class="file">NAMESPACE</span>’
file now works (again) when <code><func></code> is visible from the
namespace, e.g., imported, or in base.
</p>
</li>
<li> <p><code>getParseData(f)</code> now also works for a function defined
in the first of several ‘<span class="file"><pkg>/R/*.R</span>’ source files, thanks to
Kirill Müller's report and Duncan Murdoch's patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=16756">PR#16756</a>.
</p>
</li>
<li><p> Rd <code style="white-space: pre;">⁠\Sexpr⁠</code> macros with nested <code style="white-space: pre;">⁠#ifdef⁠</code>
conditionals were not processed.
</p>
</li>
<li><p> A non-blocking connection with non-default encoding (such as
a socket) now correctly returns from <code>readLines()</code> after new
data has arrived also when its <abbr>EOF</abbr> had been reached
previously. Thanks to Peter Meilstrup's report on R-devel and
Ivan Krylov's report and patch proposal in <a href="https://bugs.R-project.org/show_bug.cgi?id=18555">PR#18555</a>.
</p>
</li>
<li> <p><code>tools::checkRdContents()</code> failed to detect empty
argument descriptions when they spanned multiple lines,
including those generated by <code>prompt()</code>.
These cases are now noted by <code>R CMD check</code>.
</p>
</li>
<li><p> Plain-text help no longer outputs spurious colons in the
arguments list (for multi-line <code style="white-space: pre;">⁠\item⁠</code> labels in the Rd source).
</p>
</li>
<li> <p><code>kappa()</code> and <code>rcond()</code> work correctly in more cases;
<code>kappa(., norm = "2")</code> now warns that it computes the 1-norm with
(default) <code>exact = FALSE</code>; prompted by Mikael Jagan's quite
comprehensive <a href="https://bugs.R-project.org/show_bug.cgi?id=18543">PR#18543</a>.
</p>
</li>
<li><p> Rd skeletons generated by <code>prompt()</code> or
<code>promptData()</code> now use a dummy title (so <code>R CMD
build</code> works). <code>tools::checkRdContents()</code> has been updated
to detect such template leftovers, including from
<code>promptPackage()</code>.
</p>
</li>
<li><p> When S4 method dispatch fails because no method was found, the
error message now includes the signature argument names; thanks to
Michael Chirico's proposal on the R-devel list.
</p>
</li>
<li> <p><code>withAutoprint({ .. })</code> now preserves <code>srcref</code>s
previously lost, thanks to Andrew Simmons' report plus fix in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18572">PR#18572</a>.
</p>
</li>
<li> <p><code>transform.data.frame()</code> no longer adjusts names;
in particular, untransformed variables are kept as-is,
including those with syntactically invalid names (<a href="https://bugs.R-project.org/show_bug.cgi?id=17890">PR#17890</a>).
</p>
</li>
<li><p> The <code>keep.source</code> option for Rd <code style="white-space: pre;">⁠\Sexpr⁠</code> blocks is
no longer ignored.
</p>
</li>
<li><p> The <code>formula</code> methods for <code>t.test()</code> and
<code>wilcox.test()</code> now catch when <code>paired</code> is passed,
addressing <a href="https://bugs.R-project.org/show_bug.cgi?id=14359">PR#14359</a>; use <code>Pair(x1, x2) ~ 1</code> for a paired test.
</p>
</li>
<li><p> The level reported in the <code>browser</code> prompt was often too
large. It now shows the number of browser contexts on the stack.
</p>
</li>
<li><p> For <code>cbind()</code> and <code>rbind()</code>, the optional
<code>deparse.level</code> argument is now properly passed to methods,
thanks to Mikael Jagan's <a href="https://bugs.R-project.org/show_bug.cgi?id=18579">PR#18579</a> and comments there.
</p>
</li>
<li><p> Some error and warning messages for large (‘long
vector’) <code>matrix(v, nr, nc)</code> and <code>dim(m) <- d</code> are now
correct about sizes, using <code>long long</code> formatting, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18612">PR#18612</a> (and more) reported by Mikael Jagan.
</p>
</li>
<li> <p><code>readChar(useBytes = TRUE)</code> now terminates strings even when
the underlying connection uses extra spaces in the input buffer. This
fixes problems with extra garbage seen with <code>gzip</code>
connections, <a href="https://bugs.R-project.org/show_bug.cgi?id=18605">PR#18605</a>.
</p>
</li>
<li><p> Named capture in PCRE regular expressions now works also with
more than 127 named groups (<a href="https://bugs.R-project.org/show_bug.cgi?id=18588">PR#18588</a>).
</p>
</li>
<li><p> Datetime functions are now robust against long jumps when
dealing with internal time zone changes. This avoids confusing
warnings about an invalid time zone, previously triggered by turning
warnings into errors or handling them via <code>tryCatch</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=17966">PR#17966</a>,
<a href="https://bugs.R-project.org/show_bug.cgi?id=17780">PR#17780</a>).
</p>
</li>
<li><p> Datetime functions now restore even an empty <span class="env">TZ</span>
environment variable after internal time zone changes (<a href="https://bugs.R-project.org/show_bug.cgi?id=17724">PR#17724</a>).
This makes results of datetime functions with this (typically
unintentional) setting more predictable.
</p>
</li>
<li> <p><code>drop.terms(*)</code> now drops response as by default,
<code>keep.response = FALSE</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18564">PR#18564</a> thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>dummy.coef(.)</code> now also works for <code>lm()</code>-models with
<code>character</code> categorical predictor variables rather than
<code>factor</code> ones, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18635">PR#18635</a> reported by Jinsong Zhao.
</p>
</li>
<li> <p><code>formals(f) <- formals(f)</code> now also works for a function
w/o arguments and atomic <em>constant</em> <code>body(f)</code>.
</p>
</li>
<li><p> Correct <code>as.function(<invalid list>, .)</code>'s error message.
</p>
</li>
<li> <p><code>removeSource()</code> is yet more thorough in finding and
removing <code>"srcref"</code> and the other source references from parsed
<span class="rlang"><b>R</b></span> language chunks, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18638">PR#18638</a> thanks to Andrew Simmons.
</p>
</li>
<li> <p><code>dgeom()</code> is more accurate now, notably when its result is
very small, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18642">PR#18642</a> thanks to the proposal of Morten Welinder,
also improving other instances where C level <code>binom_raw(x, n, ..)</code>
has <code>x == 0</code> or <code>x== n</code>.
</p>
</li>
<li> <p><code>warning()</code> with <code>options(warn = 1)</code> has improved
output for multi-line messages.
</p>
</li>
<li> <p><code>axis.Date()</code> and <code>axis.POSIXct()</code> now respect the
<code>par("lab")</code> setting for the number of <code>pretty()</code> intervals.
</p>
</li>
<li><p> Comparisons for language objects (which are based on
deparsing) are now more careful about using accurate deparsed
results (<a href="https://bugs.R-project.org/show_bug.cgi?id=18676">PR#18676</a>).
</p>
</li>
<li><p> Plain-text help (<code>Rd2txt</code>) now correctly preserves
blank lines following single-line <code style="white-space: pre;">⁠\dontrun⁠</code> code.
</p>
</li>
<li> <p><code><POSIXlt>[*]</code> no longer sets wrong <code>"balanced"</code>
attribute, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18681">PR#18681</a> thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>str(<classed-call>)</code> now deparses the call as expected,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18684">PR#18684</a>, reported by Dave Slager.
</p>
</li>
<li><p> In Rd examples, code following the closing brace of a
<code style="white-space: pre;">⁠\dontrun⁠</code>, <code style="white-space: pre;">⁠\dontshow⁠</code> or <code style="white-space: pre;">⁠\donttest⁠</code> tag
on the same line is no longer skipped when <code>R CMD check</code>
runs the examples.
</p>
</li>
<li> <p><code>as.data.frame(matrix(*, ncol=0))</code> now gets valid
<code>names()</code> and <code>colnames()</code>; reported by Davis Vaughan
on the R-devel list.
</p>
</li>
<li><p> Internal Mathlib function <code>stirlerr(n)</code> is now almost
fully (52-bit) accurate for all <code class="reqn">n >= ~5.9</code> and more accurate
also in the range <code class="reqn">1 -- 5.9</code>. This entails small (“after
12th decimal”) changes in density functions, e.g., <code>dgamma()</code>
in <em>some</em> regions of their support. The fix was partly
prompted by Morten Welinder's <a href="https://bugs.R-project.org/show_bug.cgi?id=18640">PR#18640</a>.
</p>
</li>
<li><p> Numbers like <code>9876543.2</code> are now considered non-integer by
Mathlib internal <code>R_nonint()</code>, amending original fix of <a href="https://bugs.R-project.org/show_bug.cgi?id=15734">PR#15734</a>.
</p>
</li>
<li><p> Rd comment lines no longer cause broken paragraphs in the
rendered PDF and plain-text help. In code blocks, pure comment
lines (starting with ‘<span class="samp">⁠%⁠</span>’) no longer produce an empty line.
</p>
</li>
<li> <p><code>xtabs(Freq ~ .)</code> now consistently defaults to
<code>na.action = na.pass</code>, using <code>na.rm = FALSE</code> (added as an
argument) when summing over <code>Freq</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=17770">PR#17770</a>).
</p>
</li>
<li> <p><code>tools::testInstalledPackage()</code> is no longer silent
about failures from running examples or tests and its return code
no longer ignores failures from checking vignettes.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.3.3</h3>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>iconv()</code> now fixes up variant encoding names such as
<code>"utf8"</code> case-insensitively.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> The legacy <code>encoding = "MacRoman"</code> is deprecated in
<code>pdf()</code> and <code>postscript()</code>: support was incomplete in
earlier versions of <span class="rlang"><b>R</b></span>.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Arguments are now properly forwarded to methods on S4 generics
with ... in the middle of their formal arguments. This was
broken for the case when a method introduced an argument but did
not include ... in its own formals. Thanks to Hervé Pagès for
the report <a href="https://bugs.R-project.org/show_bug.cgi?id=18538">PR#18538</a>.
</p>
</li>
<li><p> Some invalid <code>file</code> arguments to <code>pictex()</code>,
<code>postscript()</code> and <code>xfig()</code> opened a file called
‘<span class="samp">⁠NA⁠</span>’ rather than throw an error. These included
<code>postscript(NULL)</code> (which some people expected to work like
<code>pdf(NULL)</code>).
</p>
</li>
<li><p> Passing <code>filename = NA</code> to <code>svg()</code>, <code>cairo_pdf()</code>,
<code>cairo_ps()</code> or the Cairo-based bitmap devices opened a file
called ‘<span class="samp">⁠NA⁠</span>’: it now throws an error.
</p>
</li>
<li> <p><code>quartz(file = NA)</code> opened a file called ‘<span class="samp">⁠NA⁠</span>’, including
when used as a Quartz-based bitmap device. It now gives an error.
</p>
</li>
<li> <p><code>rank(<long vector>)</code> now works, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18617">PR#18617</a>, thanks to
Ilia Kats.
</p>
</li>
<li> <p><code>seq.int()</code> did not adequately check its
<code>length.out</code> argument.
</p>
</li>
<li> <p><code>match(<POSIXct>, .)</code> is correct again for differing time
zones, ditto for <code>"POSIXlt"</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18618">PR#18618</a> reported by
Bastian Klein.
</p>
</li>
<li> <p><code>drop.terms(*, dropx = <0-length>)</code> now works, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18563">PR#18563</a> as proposed by Mikael Jagan.
</p>
</li>
<li> <p><code>drop.terms(*)</code> keeps <code> + offset(.)</code> terms when it should,
<a href="https://bugs.R-project.org/show_bug.cgi?id=18565">PR#18565</a>, and <code>drop.terms()</code> no longer makes up a response,
<a href="https://bugs.R-project.org/show_bug.cgi?id=18566">PR#18566</a>, fixing both bugs thanks to Mikael Jagan.
</p>
</li>
<li> <p><code>getS3method("t", "test")</code> no longer finds the
<code>t.test()</code> function, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18627">PR#18627</a>.
</p>
</li>
<li> <p><code>pdf()</code> and <code>postscript()</code> support for the documented
Adobe encodings <code>"Greek"</code> and <code>"Cyrilllic"</code> was missing
(although the corresponding Windows' codepages could be used).
</p>
</li>
<li><p> Computations of glyph metric information for <code>pdf()</code>
and <code>postscript()</code> did not take into account that
transliteration could replace one character by two or more (only
seen on macOS 14) and typically warned that the information was
not known.
</p>
</li>
<li> <p><code>rank(x)</code> no longer overflows during integer addition, when
computing rank average for largish but not-yet long vector <code>x</code>,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18630">PR#18630</a>, thanks to Ilia Kats.
</p>
</li>
<li> <p><code>list.files()</code> on Windows now returns also files with names
longer that 260 bytes (the Windows limit is 260 characters).
Previously, some file names particularly with ‘East Asian’
characters were omitted.
</p>
</li>
<li> <p><code>cov2cor(<0 x 0>)</code> now works, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18423">PR#18423</a> thanks to
Mikael Jagan and Elin Waring.
</p>
</li>
<li> <p><code>cov2cor(<negative diagonal>)</code> and similar now give one
warning instead of two, with better wording, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18424">PR#18424</a> thanks
to Mikael Jagan.
</p>
</li>
<li> <p><span class="pkg">tools</span><code>:: startDynamicHelp()</code> now ensures <code>port</code>
is in proper range, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18645">PR#18645</a>.
</p>
</li>
<li> <p><code>pbeta(x, a,b)</code> is correct now for <code>x</code>=0 or 1 in the
boundary cases where <code>a</code> or <code>b</code> or both are 0, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18672">PR#18672</a> thanks to Michael Fay.
</p>
</li>
<li> <p><code>pmatch(x, table)</code> for large <code>table</code>, also called for
data frame row selection, <code>dfrm[nm, ]</code>, is now interruptible,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18656">PR#18656</a>.
</p>
</li>
<li> <p><code>predict(<rank-deficient lm>, newdata=*)</code> fix computing of
<code>nbasis</code>, see Russ Lenth's comment 29 in <a href="https://bugs.R-project.org/show_bug.cgi?id=16158">PR#16158</a>.
</p>
</li>
<li><p> Added a work-around for a bug in macOS 14.3.1 and higher which
prevents R plots in the Quartz Cocoa device from updating on screen.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.3.2</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> The default initialization of the <code>"repos"</code> option
from the ‘<span class="file">repositories</span>’ file at startup can be skipped by
setting environment variable <span class="env">R_REPOSITORIES</span> to <code>NULL</code>
such that <code>getOption("repos")</code> is empty if not set elsewhere.
</p>
</li>
<li> <p><code>qr.X()</code> is now an implicit S4 generic in <span class="pkg">methods</span>.
</p>
</li>
<li> <p><code>iconv(to = "ASCII//TRANSLIT")</code> is emulated using
substitution on platforms which do not support it (notably Alpine
Linux). This should give a human-readable conversion in ASCII on
all platforms (rather than <code>NA_character_</code>).
</p>
</li>
<li> <p><code>trans3d()</code> gains options <code>continuous</code> and
<code>verbose</code> addressing the problem of possible “wrap
around” when projecting too long curves, as reported by Achim
Zeileis in <a href="https://bugs.R-project.org/show_bug.cgi?id=18537">PR#18537</a>.
</p>
</li>
<li> <p><code>tools::showNonASCII()</code> has been rewritten to work
better on macOS 14 (which has a changed implementation of
<code>iconv()</code>).
</p>
</li>
<li> <p><code>tiff(type = "quartz")</code> (the default on macOS) now warns
if <code>compression</code> is specified: it continues to be ignored.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> There is some support for building with Intel's LLVM-based
compilers on ‘<span class="samp">⁠x86_64⁠</span>’ Linux, such as (C) <code>icx</code>, (C++)
<code>ipcx</code> and (Fortran) <code>ifx</code> from oneAPI 2023.x.y.
</p>
</li>
<li><p> There is support for using LLVM's <code>flang-new</code> as the
Fortran compiler from LLVM 16.0.x (preferably 17.0.0
or later).
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> reports the use of the Fortran 90 random
number generator <code>RANDOM_NUMBER()</code> and the subroutines to
initialize it.
</p>
<p>‘Writing R Extensions’ has example code to use <span class="rlang"><b>R</b></span>'s <abbr>RNG</abbr>s
from Fortran.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>substr(x, n, L) <- cc</code> now works (more) correctly for
multibyte UTF-8 strings <code>x</code> when <code>L > nchar(x)</code>, thanks to
a report and patch by ‘Architect 95’.
</p>
</li>
<li> <p><code>contrib.url(character())</code> now returns 0-length
<code>character()</code> as documented, which also avoids spurious
warnings from <code>available.packages()</code> <abbr>et al.</abbr> in the
edge case of an empty vector of repository URLs.
</p>
</li>
<li> <p><code>readChar(., 4e8)</code> no longer fails, thanks to Kodi Arfer's
report (<a href="https://bugs.R-project.org/show_bug.cgi?id=18557">PR#18557</a>).
</p>
</li>
<li> <p><code>lapply(<list>, as.data.frame)</code> no longer warns falsely for
some base vector components.
</p>
</li>
<li><p> Communication between parent and child processes in the
<code>multicore</code> part of <code>parallel</code> could fail on platforms
that do not support an arbitrarily large payload in system
functions <code>read()</code>/<code>write()</code> on pipes (seen on macOS
where a restriction to <code>INT_MAX</code> bytes is documented, without
doing a partial read unlike Linux). The payload is now split into
1<abbr>Gb</abbr> chunks to avoid that problem. (<a href="https://bugs.R-project.org/show_bug.cgi?id=18571">PR#18571</a>)
</p>
</li>
<li> <p><code>qqplot(x,y, conf.level=.)</code> gives better confidence bounds
when <code>length(x) != length(y)</code>, thanks to Alexander Ploner's
report and patch proposal (<a href="https://bugs.R-project.org/show_bug.cgi?id=18557">PR#18557</a>).
</p>
</li>
<li> <p><code>norm(<0-length>, "2")</code> now gives zero instead of an
error, as all the other norm types, thanks to Mikael Jagan's <a href="https://bugs.R-project.org/show_bug.cgi?id=18542">PR#18542</a>.
</p>
</li>
<li><p> Build-stage Rd macros <code style="white-space: pre;">⁠\packageAuthor⁠</code> and
<code style="white-space: pre;">⁠\packageMaintainer⁠</code> now process ‘<span class="samp">⁠Authors@R⁠</span>’,
fixing ‘<span class="samp">⁠NA⁠</span>’ results when the package ‘<span class="file">DESCRIPTION</span>’ omits
‘<span class="samp">⁠Author⁠</span>’ and ‘<span class="samp">⁠Maintainer⁠</span>’ fields.
</p>
</li>
<li><p> Formatting and printing complex numbers could give things
like ‘<span class="samp">⁠0.1683-0i⁠</span>’ because of rounding error: ‘<span class="samp">⁠-0i⁠</span>’ is now
replaced by ‘<span class="samp">⁠+0i⁠</span>’.
</p>
</li>
<li> <p><code>postscript()</code> refused to accept a <code>title</code> comment
containing the letter “W” (<a href="https://bugs.R-project.org/show_bug.cgi?id=18599">PR#18599</a>).
</p>
</li>
<li> <p><code>isoreg(c(1,Inf))</code> signals an error instead of
segfaulting, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18603">PR#18603</a>.
</p>
</li>
<li> <p><code>tiff(type = "Xlib")</code> was only outputting the last page
of multi-page plots.
</p>
</li>
<li> <p><code>tools::latexToUtf8()</code> again knows about ‘<span class="samp">⁠\~{n}⁠</span>’
and other letters with tilde, fixing a regression in <span class="rlang"><b>R</b></span> 4.3.0,
and about ‘<span class="samp">⁠\^{i}⁠</span>’ as an alternative to ‘<span class="samp">⁠\^{\i}⁠</span>’
(similarly with other accents).
Furthermore, LaTeX codes for accented I letters are now
correctly converted, also fixing related mistakes in
<code>tools::encoded_text_to_latex()</code>.
</p>
</li>
<li> <p><code>tar(*, tar = "internal")</code> no longer creates out-of-spec
tar files in the very rare case of user or group names longer than
32 bytes, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17871">PR#17871</a> with thanks to Ivan Krylov.
</p>
</li>
<li><p> When using the “internal” timezone datetime code, adding
a fraction of a second no longer adds one second, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=16856">PR#16856</a>
from a patch by Ivan Krylov.
</p>
</li>
<li> <p><code>tools::checkRd()</code> no longer produces spurious notes
about “unnecessary braces” from multi-line Rd results of
<code style="white-space: pre;">⁠\Sexpr⁠</code> macros.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.3.1</h3>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The C-level API version of <span class="rlang"><b>R</b></span>'s <code>integrate()</code>,
<code>Rdqags()</code> in ‘<span class="file">Applic.h</span>’, now returns the correct number of
integrand evaluations <code>neval</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18515">PR#18515</a> reported and
diagnosed by Stephen Wade.
</p>
</li>
<li><p> The C prototypes for LAPACK calls <code>dspgv</code> and <code>dtptrs</code>
in ‘<span class="file">R_ext/Lapack.h</span>’ had one too many and one too few
character length arguments — but this has not caused any known
issues. To get the corrected prototypes, include
</p>
<pre>
#include <Rconfig.h> // for PR18534fixed
#ifdef PR18534fixed
# define usePR18534fix 1
#endif
#include <R_ext/Lapack.h>
</pre>
<p>in your C/C++ code (<a href="https://bugs.R-project.org/show_bug.cgi?id=18534">PR#18534</a>).
</p>
</li></ul>
<h4>INSTALLATION</h4>
<ul>
<li><p> Many of the checks of esoteric Internet operations and those
using unreliable external sites have been moved to a new target
that is not run by default and primarily intended for the core
developers. To run them use
</p>
<pre>
cd tests; make test-Internet-dev
</pre>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>.S3methods()</code>, typically called from <code>methods()</code>,
again marks methods from package <span class="pkg">base</span> as <code>visible</code>.
</p>
<p>Also, the visibility of non-<span class="pkg">base</span> methods is again determined by
the method's presence in <code>search()</code>.
</p>
</li>
<li> <p><code>tools::Rdiff()</code> is now more robust against invalid
strings, fixing installation tests on Windows without Rtools
installed (<a href="https://bugs.R-project.org/show_bug.cgi?id=18530">PR#18530</a>).
</p>
</li>
<li><p> Fix (new) bug in <code>hcl.colors(2, *)</code>, by Achim Zeileis (<a href="https://bugs.R-project.org/show_bug.cgi?id=18523">PR#18523</a>).
</p>
</li>
<li> <p><code>head(., <illegal>)</code> and <code>tail(..)</code> now produce
more useful <code>"Error in ...."</code> error messages, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18362">PR#18362</a>.
</p>
</li>
<li><p> Package code syntax on Windows is checked in UTF-8 when UTF-8 is
the native encoding.
</p>
</li>
<li> <p><code>na.contiguous(x)</code> now also returns the first run, when it
is at the beginning and there is a later one of the same length;
reported to R-devel, including a fix, by Georgi Boshnakov.
Further, by default, it modifies only an existing
<code>attr(*,"tsp")</code> but otherwise no longer sets one.
</p>
</li>
<li> <p><code>chol(<not pos.def>, pivot = <T|F>)</code> now gives a correct
error or warning message (depending on <code>pivot</code>), thanks to
Mikael Jagan's (<a href="https://bugs.R-project.org/show_bug.cgi?id=18541">PR#18541</a>).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.3.0</h3>
<h4>SIGNIFICANT USER-VISIBLE CHANGES</h4>
<ul>
<li><p> Calling <code>&&</code> or <code>||</code> with LHS or (if evaluated) RHS of
length greater than one is now always an error, with a report of
the form
</p>
<pre> 'length = 4' in coercion to 'logical(1)'</pre>
<p>Environment variable <span class="env">_R_CHECK_LENGTH_1_LOGIC2_</span> no longer has
any effect.
</p>
</li></ul>
<h4>NEW FEATURES</h4>
<ul>
<li><p> The included BLAS sources have been updated to those shipped
with LAPACK version 3.10.1. (This caused some platform-dependent
changes to package check output.) And then to the sources from
LAPACK version 3.11.0 (with changes only to double complex
subroutines).
</p>
</li>
<li><p> The included LAPACK sources have been updated to include the
four Fortran 90 routines rather than their Fortran 77
predecessors. This may give some different signs in <abbr>SVD</abbr>s or
eigendecompositions.. (This completes the transition to LAPACK
3.10.x begun in <span class="rlang"><b>R</b></span> 4.2.0.)
</p>
</li>
<li><p> The LAPACK sources have been updated to version 3.11.0. (No
new subroutines have been added, so this almost entirely bug
fixes: Those fixes do affect some computations with <code>NaN</code>s,
including <span class="rlang"><b>R</b></span>'s <code>NA</code>.)
</p>
</li>
<li><p> The parser now signals <em>classed</em> errors, notably in
case of the pipe operator <code>|></code>. The error object and
message now give line and column numbers, mostly as proposed and
provided by Duncan Murdoch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18328">PR#18328</a>.
</p>
</li>
<li> <p><code>toeplitz()</code> is now generalized for asymmetric cases,
with a <code>toeplitz2()</code> variant.
</p>
</li>
<li> <p><code>xy.coords()</code> and <code>xyz.coords()</code> and consequently,
e.g., <code>plot(x,y, log = "y")</code> now signal a <em>classed</em> warning
about negative values of y (where <code>log(.)</code> is <code>NA</code>).
Such a warning can be specifically suppressed or caught otherwise.
</p>
</li>
<li><p> Regular expression functions now check more thoroughly whether
their inputs are valid strings (in their encoding, e.g. in UTF-8).
</p>
</li>
<li><p> The performance of <code>grep()</code>, <code>sub()</code>,
<code>gsub()</code> and <code>strsplit()</code> has been improved,
particularly with <code>perl = TRUE</code> and <code>fixed = TRUE</code>. Use
of <code>useBytes = TRUE</code> for performance reasons should no longer
be needed and is discouraged: it may lead to incorrect results.
</p>
</li>
<li> <p><code>apropos()</code> gains an argument <code>dot_internals</code> which
is used by the completion (<code>help(rcompgen)</code>) engine to also
see <span class="pkg">base</span> internals such as <code>.POSIXct()</code>.
</p>
</li>
<li><p> Support in <code>tools::Rdiff()</code> for comparing uncompressed
PDF files is further reduced – see its help page.
</p>
</li>
<li> <p><code>qqplot(x, y, ...)</code> gains <code>conf.level</code> and
<code>conf.args</code> arguments for computing and plotting a confidence
band for the treatment function transforming the distribution of
<code>x</code> into the distribution of <code>y </code> (Switzer, 1976,
<em>Biometrika</em>). Contributed by Torsten Hothorn.
</p>
</li>
<li><p> Performance of <code>package_dependencies()</code> has been improved
for cases when the number of dependencies is large.
</p>
</li>
<li><p> Strings newly created by <code>gsub()</code>, <code>sub()</code> and
<code>strsplit()</code>, when any of the inputs is marked as <code>"bytes"</code>,
are also marked as <code>"bytes"</code>. This reduces the risk of creating
invalid strings and accidental substitution of bytes deemed invalid.
</p>
</li>
<li><p> Support for <code>readLines(encoding = "bytes")</code> has been added to
allow processing special text files byte-by-byte, without creating
invalid strings.
</p>
</li>
<li> <p><code>iconv(from = "")</code> now takes into account any declared
encoding of the input elements and uses it in preference to the native
encoding. This reduces the risk of accidental creation of invalid
strings, particularly when different elements of the input have
different encoding (including <code>"bytes"</code>).
</p>
</li>
<li><p> Package repositories in <code>getOption("repos")</code> are now
initialized from the ‘<span class="file">repositories</span>’ file when <span class="pkg">utils</span> is
loaded (if not already set, e.g., in ‘<span class="file">.Rprofile</span>’).
(From a report and patch proposal by Gabriel Becker in <a href="https://bugs.R-project.org/show_bug.cgi?id=18405">PR#18405</a>.)
</p>
</li>
<li> <p><code>compactPDF()</code> gets a <code>verbose</code> option.
</p>
</li>
<li> <p><code>type.convert()</code> and hence <code>read.table()</code> get new
option <code>tryLogical = TRUE</code> with back compatible default. When
set to false, converts <code>"F"</code> or <code>"T"</code> columns to character.
</p>
</li>
<li><p> Added new unit prefixes <code>"R"</code> and <code>"Q"</code> for
abbreviating (unrealistically large) sizes beyond <code class="reqn">10^{27}</code> in
<code>standard = "SI"</code>, thanks to Henrik Bengtsson's <a href="https://bugs.R-project.org/show_bug.cgi?id=18435">PR#18435</a>.
</p>
</li>
<li> <p><code>as.data.frame()</code>'s default method now also works fine
with atomic objects inheriting from classes such as <code>"roman"</code>,
<code>"octmode"</code> and <code>"hexmode"</code>, thus fulfilling the wish of
<a href="https://bugs.R-project.org/show_bug.cgi?id=18421">PR#18421</a>, by Benjamin Feakins.
</p>
</li>
<li><p> The <code>as.data.frame.vector()</code> utility now errors for
wrong-length <code>row.names</code>. It warned for almost six years, with
“Will be an error!”.
</p>
</li>
<li> <p><code>sessionInfo()</code> now also contains <code>La_version()</code> and
reports codepage and timezone when relevant, in both <code>print()</code>
and <code>toLatex()</code> methods which also get new option <code>tzone</code>
for displaying timezone information when <code>locale = FALSE</code>.
</p>
</li>
<li><p> New function <code>R_compiled_by()</code> reports the C and
Fortran compilers used to build <span class="rlang"><b>R</b></span>, if known.
</p>
</li>
<li> <p><code>predict(<lm>, newdata = *)</code> no longer unnecessarily creates
an <code>offset</code> of all <code>0</code>s.
</p>
</li>
<li> <p><code>solve()</code> for complex inputs now uses argument <code>tol</code>
and by default checks for ‘computational singularity’ (as
it long has done for numeric inputs).
</p>
</li>
<li> <p><code>predict(<rank-deficient lm>, newdata=*)</code> now obeys a new
argument <code>rankdeficient</code>, with new default <code>"warnif"</code>,
warning only if there are non-estimable cases in <code>newdata</code>.
Other options include <code>rankdeficient = "NA"</code>, predicting <code>NA</code>
for non-estimable <code>newdata</code> cases.
This addresses <a href="https://bugs.R-project.org/show_bug.cgi?id=15072">PR#15072</a> by Russ Lenth and is based on his original
proposal and discussions in <a href="https://bugs.R-project.org/show_bug.cgi?id=16158">PR#16158</a> also by David Firth and Elin Waring.
Still somewhat experimental.
</p>
</li>
<li> <p><code>Rgui</code> console implementation now works better with the
<code>NVDA</code> screen reader when the full blinking cursor is selected.
The underlying improvements in cursor handling may help also other
screen readers on Windows.
</p>
</li>
<li><p> The drop-field control in GraphApp can now be left with the TAB
key and all controls can be navigated in the reverse order using the
Shift+TAB key, improving accessibility of the <code>Rgui</code>
configuration editor.
</p>
</li>
<li> <p><code>qnorm(<very large negative>, log.p=TRUE)</code> is now fully
accurate (instead of to “only” minimally five digits).
</p>
</li>
<li> <p><code>demo(error.catching)</code> now also shows off
<code>withWarnings()</code> and <code>tryCatchWEMs()</code>.
</p>
</li>
<li><p> As an experimental feature the placeholder <code>_</code> can now
also be used in the <code>rhs</code> of a forward pipe <code>|></code>
expression as the first argument in an extraction call, such as
<code>_$coef</code>. More generally, it can be used as the head of a
chain of extractions, such as <code>_$coef[[2]]</code>.
</p>
</li>
<li><p> Spaces in the environment variable used to choose the <span class="rlang"><b>R</b></span>
session's temporary directory (<span class="env">TMPDIR</span>, <span class="env">TMP</span> and
<span class="env">TEMP</span> are tried in turn) are now fatal. (On Windows the
‘short path’ version of the path is tried and used if that
does not contain a space.)
</p>
</li>
<li> <p><code>all.equal.numeric()</code> gets a new optional switch
<code>giveErr</code> to return the numeric error as attribute. Relatedly,
<code>stopifnot(all.equal<some>(a, b, ..))</code> is as “smart” now,
as <code>stopifnot(all.equal(....))</code> has been already, thus allowing
customized <code>all.equal<Some>()</code> wrappers.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> on Windows is now able to work with path names longer than
260 characters when these are enabled in the system (requires at least
Windows 10 version 1607). Packages should be updated to work with
long paths as well, instead of assuming <code>PATH_MAX</code> to be the
maximum length. Custom front-ends and applications embedding <span class="rlang"><b>R</b></span> need
to update their manifests if they wish to allow this feature. See
<a href="https://blog.r-project.org/2023/03/07/path-length-limit-on-windows">https://blog.r-project.org/2023/03/07/path-length-limit-on-windows</a>
for more information.
</p>
</li>
<li> <p>‘Object not found’ and ‘Missing argument’
errors now give a more accurate error context. Patch provided by
Lionel Henry in <a href="https://bugs.R-project.org/show_bug.cgi?id=18241">PR#18241</a>.
</p>
</li>
<li><p> The <code>@</code> operator is now an S3 generic. Based on
contributions by Tomasz Kalinowski in <a href="https://bugs.R-project.org/show_bug.cgi?id=18482">PR#18482</a>.
</p>
</li>
<li><p> New generic <code>chooseOpsMethod()</code> provides a mechanism
for objects to resolve cases where two suitable methods are found
for an Ops Group Generic. This supports experimenting with
alternative object systems.
Based on contributions by Tomasz Kalinowski in <a href="https://bugs.R-project.org/show_bug.cgi?id=18484">PR#18484</a>.
</p>
</li>
<li> <p><code>inherits(x, what)</code> now accepts values other than a
simple character vector for argument <code>what</code>. A new generic,
<code>nameOfClass()</code>, is called to resolve the class name from
<code>what</code>. This supports experimenting with alternative object
systems.
Based on contributions by Tomasz Kalinowski in <a href="https://bugs.R-project.org/show_bug.cgi?id=18485">PR#18485</a>.
</p>
</li>
<li><p> Detection of BLAS/LAPACK in use (<code>sessionInfo()</code>) with
FlexiBLAS now reports the current backend.
</p>
</li>
<li><p> The <code>"data.frame"</code> method for <code>subset()</code> now warns
about extraneous arguments, typically catching the use of ‘<span class="samp">⁠=⁠</span>’
instead of ‘<span class="samp">⁠==⁠</span>’ in the <code>subset</code> expression.
</p>
</li>
<li><p> Calling <code>a:b</code> when numeric <code>a</code> or <code>b</code> is longer
than one may now be made into an error by setting environment variable
<span class="env">_R_CHECK_LENGTH_COLON_</span> to a true value, along the proposal in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18419">PR#18419</a> by Henrik Bengtsson.
</p>
</li>
<li> <p><code>density(x, weights = *)</code> now warns if automatic bandwidth
selection happens without using <code>weights</code>; new optional
<code>warnWbw</code> may suppress the warning. Prompted by Christoph
Dalitz' <a href="https://bugs.R-project.org/show_bug.cgi?id=18490">PR#18490</a> and its discussants.
</p>
</li>
<li> <p><code>rm(list = *)</code> is faster and more readable thanks to
Kevin Ushey's <a href="https://bugs.R-project.org/show_bug.cgi?id=18492">PR#18492</a>.
</p>
</li>
<li><p> The <code>plot.lm()</code> function no longer produces a normal
Q-Q plot for GLMs. Instead it plots a half-normal Q-Q plot of the
absolute value of the standardized deviance residuals.
</p>
</li>
<li><p> The <code>print()</code> method for class <code>"summary.glm"</code> no
longer shows summary statistics for the deviance residuals by
default. Its optional argument <code>show.residuals</code> can be used
to show them if required.
</p>
</li>
<li><p> The <code>tapply()</code> function now accepts a data frame as its
<code>X</code> argument, and allows <code>INDEX</code> to be a formula in that
case. <code>by.data.frame()</code> similarly allows <code>INDICES</code> to be
a formula.
</p>
</li>
<li><p> The performance of <code>df[j] <- value</code> (including for
missing <code>j</code>) and <code>write.table(df)</code> has been improved for
data frames <code>df</code> with a large number of columns. (Thanks to
Gabriel Becker's <a href="https://bugs.R-project.org/show_bug.cgi?id=18500">PR#18500</a>, <a href="https://bugs.R-project.org/show_bug.cgi?id=18503">PR#18503</a> and discussants, prompted
by a report from Toby Dylan Hocking on the R-devel mailing list.)
</p>
</li>
<li><p> The matrix multiply operator <code>%*%</code> is now an S3 generic,
belonging to new group generic <code>matrixOps</code>.
From Tomasz Kalinowski's contribution in <a href="https://bugs.R-project.org/show_bug.cgi?id=18483">PR#18483</a>.
</p>
</li>
<li><p> New function <code>array2DF()</code> to convert arrays to data
frames, particularly useful for the list arrays created by
<code>tapply()</code>.
</p>
</li></ul>
<h4>DATES and TIMES</h4>
<ul>
<li><p> On platforms where (non-UTC) datetimes before 1902 (or
before 1900 as with system functions on recent macOS) are guessed
by extrapolating time zones from 1902-2037, there is a warning at
the first use of extrapolation in a session. (As all time zones
post 2037 are extrapolation, we do not warn on those.)
</p>
</li>
<li><p> (Platforms using <span class="option">--with-internal-tzone</span>, including
Windows and by default macOS). How years are printed in dates or
date-times can be controlled by environment variable
<span class="env">R_PAD_YEARS_BY_ZERO</span>. The default remains to pad to 4 digits
by zeroes, but setting value ‘<span class="samp">⁠no⁠</span>’ gives no padding (as used
by default by <code>glibc</code>).
</p>
</li>
<li> <p><code>strftime()</code> tries harder to determine the offset for
the <code>"%z"</code> format, and succeeds on the mainstream <span class="rlang"><b>R</b></span> platforms.
</p>
</li>
<li> <p><code>strftime()</code> has a limit of 2048 bytes on the string
produced – attempting to exceed this is an error. (Previously it
silently truncated at 255 bytes.)
</p>
</li>
<li> <p><code>sessionInfo()</code> records (and by default prints) the
system time zone as part of the locale information. Also, the
source (system/internal) of the date-time conversion and printing
functions.
</p>
</li>
<li><p> Objects of class <code>"POSIXlt"</code> created in this version of
<span class="rlang"><b>R</b></span> always have 11 components: component <code>zone</code> is always set,
and component <code>gmtoff</code> is set for times in UTC and usually
set on the (almost all) platforms which have C-level support,
otherwise is <code>NA</code>.
</p>
</li>
<li><p> There are comprehensive validity checks on the structure of
objects of class <code>"POSIXlt"</code> when converting (including
formatting and printing). (This avoids mis-conversions of
hand-crafted objects.)
</p>
</li>
<li><p> There is some support for using the native date-time
routines on macOS: this is only viable on recent versions
(e.g. 12.6 and 13) and does get wrong some historical changes
(before 1900, during WWII). Use of <span class="option">--with-internal-tzone</span>
remains the default.
</p>
</li>
<li> <p><code>as.POSIXct(<numeric>)</code> and <code>as.POSIXlt(.)</code> (without
specifying <code>origin</code>) now work.
So does <code>as.Date(<numeric>)</code>.
</p>
</li>
<li> <p><code>as.Date.POSIXct(., tz)</code> now treats several <code>tz</code>
values, notably <code>"GMT"</code> as equivalent to <code>"UTC"</code>, proposed
and improved by Michael Chirico and Joshua Ulrich in <a href="https://bugs.R-project.org/show_bug.cgi?id=17674">PR#17674</a>.
</p>
</li>
<li><p> Experimental <code>balancePOSIXlt()</code> utility allows using
“ragged” and or out-of-range <code>"POSIXlt"</code> objects more
correctly, e.g., in subsetting and subassignments. Such objects
are now documented.
Complemented by the low-level <code>unCfillPOSIXlt()</code> utility.
</p>
<p>More experimentally, a <code>"POSIXlt"</code> object may have an attribute
<code>"balanced"</code> indicating if it is known to be filled or fully
balanced.
</p>
</li>
<li><p> Functions <code>axis.Date()</code> and <code>axis.POSIXct()</code> are
rewritten to gain better default tick locations and better default
formats via the corresponding <code>pretty()</code> methods.
Thanks to Swetlana Herbrandt.
</p>
</li>
<li><p> The mapping of Windows' names for time zones to <abbr>IANA</abbr>'s
‘Olson’ names has been updated. When ICU is available (it is
by default), it is used to get a mapping for the current region set in
Windows. This can be overridden by setting environment variable
<span class="env">TZ</span> to the desired Olson name — see <code>OlsonNames()</code> for
those currently available.
</p>
</li></ul>
<h4>GRAPHICS</h4>
<ul>
<li><p> The graphics engine version, <code>R_GE_version</code>, has been
bumped to <code>16</code> and so packages that provide graphics devices
should be reinstalled.
</p>
</li>
<li><p> The <span class="pkg">grDevices</span> and <span class="pkg">grid</span> packages have new
functions for rendering typeset glyphs, primarily:
<code>grDevices::glyphInfo()</code> and <code>grid::grid.glyph()</code>.
</p>
<p>Rendering of typeset glyphs is only supported so far on the
Cairo-based graphics devices and on the <code>pdf()</code> and
<code>quartz()</code> devices.
</p>
</li>
<li><p> The defined behaviour for <code>"clear"</code> and <code>"source"</code>
compositing operators (via <code>grid::grid.group()</code>) has been
changed (to align better with
simple interpretation of original Porter-Duff definitions).
</p>
</li>
<li><p> Support for gradients, patterns, clipping paths, masks,
groups, compositing operators, and affine transformations has
been added to the <code>quartz()</code> device.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> A system installation of generic LAPACK 3.10.0 or later will
be preferred to the version in the <span class="rlang"><b>R</b></span> sources.
</p>
<p><code>configure</code> option <span class="option">--with-lapack=no</span> (equivalently
<span class="option">--without-lapack</span>) forces compilation of the internal
LAPACK sources.
</p>
<p>If <span class="option">--with-lapack</span> is not specified, a system
<code>liblapack</code> is looked for and used if it reports version
3.10.0 or later and does not contain BLAS routines.
</p>
<p>Packages using LAPACK will need to be reinstalled if this changes
to using an external library.
</p>
</li>
<li><p> On ‘<span class="samp">⁠aarch64⁠</span>’ Linux platforms using GCC,
<code>configure</code> now defaults to <span class="option">-fPIC</span> (instead of
<span class="option">-fpic</span>), as desired in <a href="https://bugs.R-project.org/show_bug.cgi?id=18326">PR#18326</a>.
</p>
</li>
<li> <p><code>configure</code> now checks conversion of datetimes
between <code>POSIXlt</code> and <code>POSIXct</code> around year 2020. Failure
(which has been seen on platforms missing <code>tzdata</code>) is fatal.
</p>
</li>
<li><p> If <code>configure</code> option
<span class="option">--with-valgrind-instrumentation</span> is given value <code>1</code>
or <code>2</code>, option <span class="option">--with-system-valgrind-headers</span> is now
the default and ignored (with a warning). It is highly
recommended that the system headers are installed alongside
<code>valgrind</code>: they are part of its packaging on some Linux
distributions and packaged separately (e.g. in the
‘<span class="file">valgrind-devel</span>’ RPM) on others. <code>configure</code> will
give a warning if they are not found.
</p>
<p>The system headers will be required in a future release of <span class="rlang"><b>R</b></span>
to build with <code>valgrind</code> instrumentation.
</p>
</li>
<li> <p><code>libcurl</code> 8.x is now accepted by <code>configure</code>:
despite a change in major version number it changes neither API
nor <abbr>ABI</abbr>.
</p>
</li></ul>
<h4>INSTALLATION on WINDOWS</h4>
<ul>
<li><p> The makefiles and installer scripts for Windows have been
tailored to Rtools43, an update of the Rtools42
toolchain. It is based on GCC 12 and newer versions of
MinGW-W64, binutils and libraries. At this time R-devel can still
be built using Rtools42 without changes, but when R-devel is
installed via the installer, it will by default look for Rtools43.
</p>
</li>
<li><p> Old make targets <code>rsync-extsoft</code> and 32-bit ones that
are no longer needed have been removed.
</p>
</li>
<li><p> Default builds (including for packages) no longer select
C99. Thus the C standard used is the default for the compiler,
which for the toolchain in ‘<span class="file">Rtools43</span>’ is C17. (This is
consistent with Unix builds.)
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> The default C++ standard has been changed to C++17 where
available (which it is on all currently checked platforms): if not
C++14 or C++11 is used if available otherwise C++ is not supported.
</p>
</li>
<li> <p><code>USE_FC_LEN_T</code> is the default: this uses the
correct (compiler-dependent) prototypes for Fortran BLAS/LAPACK
routines called from C/C++, and requires adjustment of many such
calls – see ‘Writing R Extensions’ §6.6.1.
</p>
</li>
<li><p> There is initial support for C++23 as several compilers are
now supporting <code>-std=c++23</code> or <code>-std=c++2b</code> or similar.
As for C++20, there no additional <code>configure</code> checks for
C++23 features beyond a check that the compiler reports a
<code>__cplusplus</code> value greater than that in the C++20 standard.
C++ feature tests should be used.
</p>
</li>
<li><p> There is support for a package to indicate the version of
the C standard which should be used to compile it, and for the
installing user to specify this. In most cases <span class="rlang"><b>R</b></span> defaults to the
C compiler's default standard which is C17 (a ‘bug-fix’ of C11) –
earlier versions of <span class="rlang"><b>R</b></span> or compilers may have defaulted to C99.
</p>
<p>Current options are:
</p>
<dl>
<dt>USE_C17</dt><dd><p>Use a standard that is at most C17. The
intention is to allow legacy packages to still be installed when
later C standards become the default, including packages
using new keywords as identifiers or with K&R-style function
declarations. This will use C17 if available, falling back
to C11.
</p>
</dd>
<dt>USE_C90</dt><dd><p>Use the C90 (aka C89) standard. (As that standard
did not require compilers to identify that version, all we can
verify is that the compiler does not claim to be using a later
standard. It may accept C99 features – for example
<code>clang</code> accepts // to make comments.)</p>
</dd>
<dt>USE_C99</dt><dd><p>Use the C99 standard. This should be rarely
needed – it avoids the few new features of C11/C17 which can
be useful if a package assumes them if C17 is specified and
they are not implemented.</p>
</dd>
<dt>USE_C23</dt><dd><p>Use C23 (or in future, later). Compiler/library
support for C23 is still being implemented, but LLVM clang
from 15.0.0 and GCC from 13 have quite extensive support.
</p>
</dd>
</dl>
<p>These can be specified as part of the ‘<span class="samp">⁠SystemRequirements⁠</span>’
field in the package's ‘<span class="file">DESCRIPTION</span>’ file or <em>via</em>
options <span class="option">--use-C17</span> and so on of <code>R CMD INSTALL</code>
and <code>R CMD SHLIB</code>.
</p>
<p>For further details see “Writing R Extensions” §1.2.5.
</p>
</li>
<li><p> (Windows) A ‘<span class="file">src/Makefile.ucrt</span>’ or
‘<span class="file">src/Makefile.win</span>’ file is now included after
‘<span class="file"><var>R_HOME</var>/etc<var>R_ARCH</var>/Makeconf</span>’ and so no longer
needs to include that file itself. Installation of a package with
such a file now uses a site ‘<span class="file">Makevars</span>’ file in the same way
as a package with a ‘<span class="file">src/Makevars.win</span>’ file would.
</p>
</li>
<li> <p><code>configure</code> is now passed crucial variables such as
<span class="env">CC</span> and <span class="env">CFLAGS</span> in its environment, as many packages
were not setting them (as documented in ‘Writing R
Extensions’ §1.2).
</p>
<p>This has most effect where <code>configure</code> is used to compile
parts of the package – most often by <code>cmake</code> or
<code>libtool</code> which obfuscate the actual compile commands
used.
</p>
<p>Also used for <code>configure.win</code> and <code>configure.ucrt</code>
on Windows.
</p>
</li></ul>
<h4>FORTRAN FLAGS</h4>
<ul>
<li><p> The flag <span class="option">-fno-optimize-sibling-calls</span> is no longer
forced for <code>gfortran</code> 7 and later. It should no longer be
needed now using ‘hidden’ character-length arguments when
calling BLAS/LAPACK routines from C/C++ is the default even for
packages. (Unless perhaps packages call Fortran code from C/C++
without using <span class="rlang"><b>R</b></span>'s headers and without allowing for these
arguments.)
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The deprecated S-compatibility macros <code>DOUBLE_*</code> in
‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’) have been
removed.
</p>
</li>
<li><p> The deprecated legacy typedefs of <code>Sint</code> and
<code>Sfloat</code> in header ‘<span class="file">R.h</span>’ are no longer defined, and that
header no longer includes header ‘<span class="file">limits.h</span>’ from C nor
‘<span class="file">climits</span>’ from C++.
</p>
</li>
<li><p> New macro <code>CAD5R()</code> is provided in ‘<span class="file">Rinternals.h</span>’
and used in a few places in the <span class="rlang"><b>R</b></span> sources.
</p>
</li>
<li> <p>ALTREP now supports <code>VECSXP</code> vectors. Contributed by
Gabor Csardi in <a href="https://bugs.R-project.org/show_bug.cgi?id=17620">PR#17620</a>.
</p>
</li>
<li><p> The <code>Rcomplex</code> definition (in header
‘<span class="file">R_ext/Complex.h</span>’) has been extended to prevent possible
mis-compilation when interfacing with Fortran (<a href="https://bugs.R-project.org/show_bug.cgi?id=18430">PR#18430</a>). The
new definition causes compiler warnings with static initializers
such as <code>{1, 2}</code>, which can be changed to <code>{.r=1,
.i=2}</code>.
</p>
<p>Using the new definition from C++ depends on compiler extensions
supporting C features that have not been incorporated into the C++
standards but are available in <code>g++</code> and
<code>clang++</code>: this may result in C++ compiler warnings but
these have been worked around for recent versions of common
compilers (GCC, Apple/LLVM clang, Intel).
</p>
<p>It is intended to change the inclusion of header
‘<span class="file">R_ext/Complex.h</span>’ by other <span class="rlang"><b>R</b></span> headers, so C/C++ code files
which make use of <code>Rcomplex</code> should include that header
explicitly.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> does more checking of package ‘<span class="file">.Rd</span>’
files, warning about invalid email addresses and (some) invalid
<abbr>URI</abbr>s and noting empty ‘<span class="samp">⁠\item⁠</span>’ labels in description lists.
</p>
</li>
<li> <p><code>R CMD check</code> now also reports problems when reading
package news in <abbr>md</abbr> (file ‘<span class="file">NEWS.md</span>’) and (optionally) plain
text (file ‘<span class="file">NEWS</span>’) formats.
</p>
</li>
<li> <p><span class="env">_R_CHECK_TIMINGS_</span> defaults to a value from the
environment even for <code>R CMD check --as-cran</code>; this allows
for exceptionally fast or slow platforms.
</p>
<p>It now applies to checking PDF and HTML versions of the manuals,
and ‘checking CRAN incoming feasibility’.
</p>
</li>
<li> <p><code>R CMD check</code> can optionally (but included in
<span class="option">--as-cran</span>) check whether HTML math rendering <em>via</em>
KaTeX works for the package ‘<span class="file">.Rd</span>’ files.
</p>
</li>
<li><p> Non-interactive debugger invocations can be trapped by
setting the environment variable
<span class="env">_R_CHECK_BROWSER_NONINTERACTIVE_</span> to a true value. This is
enabled by <code>R CMD check --as-cran</code> to detect the use of
leftover <code>browser()</code> statements in the package.
</p>
</li>
<li><p> The use of <code>sprintf</code> and <code>vsprintf</code> from C/C++ has
been deprecated in macOS 13 and is a known security risk.
<code>R CMD check</code> now reports (on all platforms) if their use
is found in compiled code: replace by <code>snprintf</code> or
<code>vsnprintf</code> respectively. [<b>NB:</b> whether such calls get
compiled into the package is platform-dependent.]
</p>
</li>
<li><p> Where recorded at installation, <code>R CMD check</code>
reports the C and Fortran compilers used to build <span class="rlang"><b>R</b></span>.
</p>
<p>It reports the OS in use (if known, as given by
<code>osVersion</code>) as well as that <span class="rlang"><b>R</b></span> was built for.
</p>
<p>It notes if a C++ standard was specified which is older than the
current default: many packages have used C++11 to mean ‘not
C++98’ — as C++11 is the minimum supported since <span class="rlang"><b>R</b></span> 4.0.0, that
specification can probably be removed.
</p>
</li>
<li> <p><code>R CMD INSTALL</code> reports the compilers (and on
macOS, the <abbr>SDK</abbr>) used, and this is copied to the output of
<code>R CMD check</code>.
</p>
<p>Where a C++ standard is specified, it is reported.
</p>
</li>
<li> <p><code>R CMD check</code>'s
‘checking compilation flags in Makevars’
has been relaxed to accept the use of flags such as
<span class="option">-std=f2008</span> in ‘<span class="samp">⁠PKG_FFLAGS⁠</span>’.
</p>
</li>
<li> <p><code>tools::buildVignettes()</code> has a new argument <code>skip</code>,
which is used by <code>R CMD check</code> to skip (and note)
vignettes with unavailable ‘<span class="samp">⁠\VignetteDepends⁠</span>’ (<a href="https://bugs.R-project.org/show_bug.cgi?id=18318">PR#18318</a>).
</p>
</li>
<li><p> New generic <code>.AtNames()</code> added to enable class-specific
completions after <code>@</code>. The formerly internal function
<code>findMatches()</code> is now exported, mainly for use in methods
for <code>.DollarNames()</code> and <code>.AtNames()</code>.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li> <p><code>default.stringsAsFactors()</code> is defunct.
</p>
</li>
<li><p> Calling <code>as.data.frame.<class>()</code> directly (for 12 atomic
classes) is going to be formally deprecated, currently activated by
setting the environment variable
<span class="env">_R_CHECK_AS_DATA_FRAME_EXPLICIT_METHOD_</span> to non-empty, which
also happens in <code>R CMD check --as-cran</code>.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Hashed <code>environment</code>s with sizes less than 5 can now grow.
(Reported to R-devel by Duncan Garmonsway.)
</p>
</li>
<li> <p><code>as.character(<Rd>, deparse = TRUE)</code> failed to
re-escape curly braces in LaTeX-like text.
(Reported by Hadley Wickham in <a href="https://bugs.R-project.org/show_bug.cgi?id=18324">PR#18324</a>.)
</p>
</li>
<li> <p><code>library()</code> now passes its <code>lib.loc</code> argument when
requiring <code>Depends</code> packages; reported (with fix) in <a href="https://bugs.R-project.org/show_bug.cgi?id=18331">PR#18331</a>
by Mikael Jagan.
</p>
</li>
<li> <p><code>R CMD Stangle</code>: improved message about ‘Output’
files.
</p>
</li>
<li> <p><code>head(x, n)</code> and <code>tail(x, n)</code> now signal an error if
<code>n</code> is not numeric, instead of incidentally “working”
sometimes returning all of <code>x</code>. Reported and discussed by Colin
Fay, in <a href="https://bugs.R-project.org/show_bug.cgi?id=18357">PR#18357</a>.
</p>
</li>
<li><p> The <code>"lm"</code> method for <code>summary()</code> now gives the
correct F-statistic when the model contains an offset. Reported in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18008">PR#18008</a>.
</p>
</li>
<li> <p><code>C()</code> and <code>`contrasts<-`()</code> now preserve factor
level names when given a function object (as opposed a function
name which did preserve names). Reported in <a href="https://bugs.R-project.org/show_bug.cgi?id=17616">PR#17616</a>.
</p>
</li>
<li> <p><code>c(a = 1, 2)[[]]</code> no longer matches <code>2</code> but rather
signals a <em>classed</em> error. Reported and analysed by Davis Vaughan in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18367">PR#18367</a>, a duplicate of <a href="https://bugs.R-project.org/show_bug.cgi?id=18004">PR#18004</a>, by Jan Meis <abbr>et al.</abbr>
For consistency, <code>NULL[[]]</code> is also erroneous now.
<code>x[[]] <- v</code> gives an error of the same class <code>"MissingSubscriptError"</code>.
</p>
</li>
<li><p> The <code>relist()</code> function of <span class="pkg">utils</span> now supports
<code>NULL</code> elements in the skeleton (<a href="https://bugs.R-project.org/show_bug.cgi?id=15854">PR#15854</a>).
</p>
</li>
<li> <p><code>ordered(levels = *)</code> (missing <code>x</code>) now works
analogously to <code>factor(, ordered=TRUE)</code>; reported (with fix)
by Achim Zeileis in <a href="https://bugs.R-project.org/show_bug.cgi?id=18389">PR#18389</a>.
</p>
</li>
<li><p> User-defined Rd macro definitions can now span multiple lines,
thanks to a patch from Duncan Murdoch. Previously, the Rd parser
silently ignored everything after the first line.
</p>
</li>
<li><p> Plain-text help (<code>tools::Rd2txt()</code>) now preserves an
initial blank line for text following description list items.
</p>
</li>
<li> <p><code>tools::Rd2HTML()</code> and <code>tools::Rd2latex()</code> no
longer split <code style="white-space: pre;">⁠\arguments⁠</code> and <code style="white-space: pre;">⁠\value⁠</code> lists at Rd comments.
</p>
</li>
<li> <p><code>tools::Rd2latex()</code> now correctly handles optional text
outside <code style="white-space: pre;">⁠\item⁠</code>s of argument lists as well as bracketed text
at the beginning of sections, e.g., <code style="white-space: pre;">⁠\value{[NULL]}⁠</code>.
</p>
</li>
<li> <p><code>as.character(<POSIXt>)</code> now behaves more in line with the
methods for atomic vectors such as numbers, and is no longer
influenced by <code>options()</code>.
Ditto for <code>as.character(<Date>)</code>.
The <code>as.character()</code> method gets arguments <code>digits</code> and <code>OutDec</code>
with defaults <em>not</em> depending on <code>options()</code>.
Use of <code>as.character(*, format = .)</code> now warns.
</p>
</li>
<li><p> Similarly, the <code>as.character.hexmode()</code> and
<code>*.octmode()</code> methods also behave as <code>good citizen</code>
methods and back compatibility option <code>keepStr = TRUE</code>.
</p>
</li>
<li><p> The <code>as.POSIXlt(<POSIXlt>)</code> and
<code>as.POSIXct(<POSIXct>)</code> default methods now do obey their
<code>tz</code> argument, also in this case.
</p>
</li>
<li> <p><code>as.POSIXlt(<Date>)</code> now does apply a <code>tz</code> (time
zone) argument, as does <code>as.POSIXct()</code>; partly suggested by
Roland Fuß on the R-devel mailing list.
</p>
</li>
<li> <p><code>as.Date.POSIXlt(x)</code> now also works when the list components
are of unequal length, aka “partially filled” or “ragged”.
</p>
</li>
<li> <p><code>expand.model.frame()</code> looked up variables in the wrong
environment when applied to models fitted without <code>data</code>.
Reported in <a href="https://bugs.R-project.org/show_bug.cgi?id=18414">PR#18414</a>.
</p>
</li>
<li> <p><code>time()</code> now (also) uses the
<code>ts.eps = getOption("ts.eps")</code>
argument and thus by default rounds values very close to the start (or
end) of a year. Based on a proposal by Andreï V. Kostyrka on R-help.
</p>
</li>
<li><p> Printing of a <code>factanal()</code> result with just one factor and
<code>sort = TRUE</code> now works regularly, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17863">PR#17863</a> by Timothy
Bates, thanks to the ‘R Contributors’ working group.
</p>
</li>
<li><p> Printing 0-length objects of class
<code>"factor"</code>, <code>"roman"</code>, <code>"hexmode"</code>,
<code>"octmode"</code>, <code>"person"</code>, <code>"bibentry"</code>, or
<code>"citation"</code> now prints something better, one
of which fixes <a href="https://bugs.R-project.org/show_bug.cgi?id=18422">PR#18422</a>, reported by Benjamin Feakins.
</p>
</li>
<li> <p><code>Sys.timezone()</code> queries <code>timedatectl</code> only if
<code>systemd</code> is loaded; addressing a report by Jan Gorecki in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17421">PR#17421</a>.
</p>
</li>
<li><p> The formula method of <code>cor.test()</code> had scoping problems
when <code>environment(formula)</code> was not the calling environment;
reported with a patch proposal by Mao Kobayashi in <a href="https://bugs.R-project.org/show_bug.cgi?id=18439">PR#18439</a>.
</p>
</li>
<li> <p><code>attach()</code> of an environment with active bindings now
preserves the active bindings. Reported by Kevin Ushey in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18425">PR#18425</a>.
</p>
</li>
<li><p> BLAS detection now works also with system-provided libraries not
available as regular files. This fixes detection of the Accelerate
framework on macOS since Big Sur. Reported by David Novgorodsky.
</p>
</li>
<li> <p><code>download.file()</code> gives a helpful error message in case of
an invalid <code>download.file.method</code> option, thanks to Colin Fay's
report in <a href="https://bugs.R-project.org/show_bug.cgi?id=18455">PR#18455</a>.
</p>
</li>
<li><p> Sporadic crashes of <code>Rterm</code> when using completion have
been fixed.
</p>
</li>
<li> <p><code>Rprof()</code> is now more reliable. A livelock in thread
initialization with too short sampling interval has been fixed on
macOS. A deadlock in using the C runtime has been fixed on Windows.
A potential deadlock has been prevented on Unix.
</p>
</li>
<li><p> Cursor placement in <code>Rgui</code> now works even after a
fixed-width font is selected.
</p>
</li>
<li><p> Mandatory options (<code>options()</code>) are now set on startup so
that saving and restoring them always works (<a href="https://bugs.R-project.org/show_bug.cgi?id=18372">PR#18372</a>).
</p>
</li>
<li><p> Package installation, <code>R CMD INSTALL</code> or <code>install.packages(*)</code>,
now parses each of the ‘<span class="file"><pkg>/R/*.R</span>’ files individually instead
of first concatenating and then <code>parse()</code>ing the large resulting
file. This allows parser or syntax errors to be diagnosed with
correct file names and line numbers, thanks to Simon Dedman's report
and Bill Dunlap's patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=17859">PR#17859</a>.
</p>
<p>This <em>does</em> require syntactically self contained R source
files now, fixing another inadvertent bug.
</p>
</li>
<li> <p><code>predict.lm(<model with offset>)</code> now finds the offset in
the correct environment, thanks to André Gillibert's report and patch
in <a href="https://bugs.R-project.org/show_bug.cgi?id=18456">PR#18456</a>.
</p>
</li>
<li> <p><code>getInitial(<formula>)</code> now finds the <code>selfStart</code>
model in the correct environment.
(Reported by Ivan Krylov in <a href="https://bugs.R-project.org/show_bug.cgi?id=18368">PR#18368</a>.)
</p>
</li>
<li><p> Fix for possible segfault when using recently-added graphics
features, such as gradients, clipping paths, masks, and groups with
<code>pdf(file=NULL)</code>.
</p>
</li>
<li> <p><code>class(m) <- class(m)</code> no longer changes a matrix <code>m</code>
by adding a class <em>attribute</em>.
</p>
</li>
<li> <p><code>packageDate(pkg)</code> now only warns once if there is no <code>pkg</code>.
</p>
</li>
<li><p> When <code>ts()</code> creates a multivariate time series,
<code>"mts"</code>, it also inherits from <code>"array"</code> now, and
<code>is.mts()</code> is documented <em>and</em> stricter.
</p>
</li>
<li> <p><code>Rd2txt()</code> now preserves line breaks of <code style="white-space: pre;">⁠\verb⁠</code> Rd
content and from duplicated <code style="white-space: pre;">⁠\cr⁠</code>. The former also
fixes the rendering of verbatim output from Rd
<code style="white-space: pre;">⁠\Sexpr⁠</code> in plain-text help.
</p>
</li>
<li> <p><code>uniroot(f, interval)</code> should no longer wrongly converge
<em>outside</em> the interval in some cases where <code>abs(f(x)) == Inf</code>
for an <code>x</code> at the interval boundary, thanks to posts by
Ben Bolker and Serguei Sokol on R-devel.
</p>
</li>
<li><p> Vectorized alpha handling in palette functions such as in
<code>gray()</code>, <code>rainbow()</code>, or <code>hcl.colors()</code> works
correctly now, thanks to Achim Zeileis' report and patch in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18476">PR#18476</a>.
</p>
</li>
<li><p> Formatting and <code>print()</code>ing of <code>bibentry</code> objects has
dropped the deprecated <code>citation.bibtex.max</code> argument, such that
the <code>bibtex</code> argument's default for <code>print.bibentry()</code>
depends directly on the <code>citation.bibtex.max</code> option, whereas in
<code>format.bibentry()</code> the option no longer applies.
</p>
</li>
<li><p> Attempting to use a character string naming a foreign
function entry point in a foreign function call in a package will
now signal an error if the packages has called
<code>R_forceSymbols</code> to specify that symbols must be
used.
</p>
</li>
<li><p> An error in <code>table()</code> could permanently set
<code>options(warn=2)</code> promoting all subsequent warnings to errors.
</p>
</li>
<li><p> The <code>sigma()</code> function gave misleading results for
binary GLMs. A new method for objects of class <code>"glm"</code> returns the
square root of the estimate of the dispersion parameter using the
same calculation as <code>summary.glm()</code>.
</p>
</li>
<li> <p><code>bs()</code> and <code>ns()</code> in the (typical) case of automatic
knot construction, when some of the supposedly inner knots coincide
with boundary knots, now moves them inside (with a warning), building
on <a href="https://bugs.R-project.org/show_bug.cgi?id=18442">PR#18442</a> by Ben Bolker.
</p>
</li>
<li> <p><code>R CMD</code> on Windows now skips the site profile with
<span class="option">--no-site-file</span> and <span class="option">--vanilla</span> even when
<span class="env">R_PROFILE</span> is set (<a href="https://bugs.R-project.org/show_bug.cgi?id=18512">PR#18512</a>, from Kevin Ushey).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.2.3</h3>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The definition of <code>DL_FUNC</code> in ‘<span class="file">R_ext/Rdynload.h</span>’
has been changed to be fully C-compliant. This means that
functions loaded <em>via</em> for example <code>R_GetCCallable</code> need
to be cast to an appropriate type if they have any arguments.
</p>
</li>
<li> <p><code>.Machine</code> has a new element <code>sizeof.time_t</code> to
identify old systems with a 32-bit type and hence a limited range
of date-times (and limited support for dates millions of years
from present).
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> (Windows) The default C++ standard had accidentally been
left at C++11 when it was changed to C++14 on Unix.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> As <code>"POSIXlt"</code> objects may be “partially filled”
and their list components meant to be recycled, <code>length()</code> now
is the length of the longest component.
</p>
</li>
<li> <p><code>as.POSIXlt.Date()</code> could underflow for dates in the
far past (more than half a million years <abbr>BCE</abbr>).
</p>
</li>
<li> <p><code>as.Date.POSIXlt(x)</code> would return <code>"1970-01-01"</code>
instead of <code>NA</code> in R 4.2.2, e.g., for </p>
<pre>
x <- as.POSIXlt(c("2019-01-30","2001-1-1"))
x$mon <- c(0L, NA); as.Date(x)</pre>
</li>
<li> <p><code>R CMD check</code> failed to apply enabled
<span class="env">_R_CHECK_SUGGESTS_ONLY_</span> to examples and vignettes
(regression in <span class="rlang"><b>R</b></span> 4.2.0).
</p>
</li>
<li> <p><code>R CMD check</code> did not re-build vignettes in
separate processes by default (regression in <span class="rlang"><b>R</b></span> 4.2.0).
</p>
</li>
<li><p> Running examples from HTML documentation now restores
previous <span class="pkg">knitr</span> settings and options (<a href="https://bugs.R-project.org/show_bug.cgi?id=18420">PR#18420</a>).
</p>
</li>
<li><p> Quartz: fonts are now located using Core Graphics API
instead of deprecated ATS which is no longer supported in the
macOS 13 <abbr>SDK</abbr> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18426">PR#18426</a>). This also addresses an issue where
the currently used font in the Quartz device context was not
correctly retained.
</p>
</li>
<li><p> (Windows) Math symbols in text drawing functions are again
rendered correctly (<a href="https://bugs.R-project.org/show_bug.cgi?id=18440">PR#18440</a>). This fixes a regression in <span class="rlang"><b>R</b></span> 4.2.1
caused by a fix in <a href="https://bugs.R-project.org/show_bug.cgi?id=18382">PR#18382</a> which uncovered an issue in GraphApp
due to which the symbol <abbr>charset</abbr> was not used with TT Symbol font face.
</p>
</li>
<li><p> (Windows) Installing a package with a
‘<span class="file">src/Makefile.{win,ucrt}</span>’ file includes
‘<span class="file">~/.R/Makevars.win64</span>’ in the search for user makevars, as
documented in “R Installation and Administration” and done
for packages with a ‘<span class="file">src/Makevars.{win,ucrt}</span>’ file.
</p>
</li>
<li> <p><code>format(<POSIXlt_w/_unbalanced_sec>, "....%OS<n>")</code> with
<code class="reqn">n > 0</code> no longer accidentally uses the unbalanced seconds,
thanks to Suharto Anggono's report (including patch) in <a href="https://bugs.R-project.org/show_bug.cgi?id=18448">PR#18448</a>.
</p>
</li>
<li> <p><code>solve.default(a, b)</code> works around issues with
some versions of LAPACK when <code>a</code> contains <code>NA</code> or
<code>NaN</code> values.
</p>
</li>
<li><p> When <code>UseMethod()</code> cannot dispatch, it no longer segfaults
producing the error message in case of a long <code>class()</code>, thanks to
Joris Vankerschaver's report (including patch) in <a href="https://bugs.R-project.org/show_bug.cgi?id=18447">PR#18447</a>.
</p>
</li>
<li><p> When <code>example(foo, ..)</code> produces graphics on an interactive
device it needs to open itself, it now leaves <code>devAskNewPage()</code>
unchanged even when it was <code>FALSE</code>, thus fixing a 14 years old
‘FIXME’.
</p>
</li>
<li> <p><code>packageDescription()</code> again catches errors from encoding
conversions. This also fixes broken <code>packageVersion()</code> in
C locale on systems where <code>iconv</code> does not support transliteration.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.2.2</h3>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>tools::Rdiff(useDiff = TRUE)</code> checks for the presence
of an external <code>diff</code> command and switches to
<code>useDiff = FALSE</code> if none is found.
This allows <code>R CMD Rdiff</code> to always work.
</p>
</li>
<li><p> On Windows, environment variable
<span class="env">R_LIBCURL_SSL_REVOKE_BEST_EFFORT</span> can be used to switch to
only ‘best-effort’ <abbr>SSL</abbr> certificate revocation checks with
the default <code>"libcurl"</code> download method. This reduces
security, but may be needed for downloads to work with <abbr>MITM</abbr>
proxies (<a href="https://bugs.R-project.org/show_bug.cgi?id=18379">PR#18379</a>).
</p>
</li>
<li><p> (macOS) The run-time check for libraries from XQuartz for
X11 and Tcl/Tk no longer uses <code>otool</code> from the Apple
Developer Tools (<a href="https://bugs.R-project.org/show_bug.cgi?id=18400">PR#18400</a>).
</p>
</li>
<li><p> The LaTeX style for producing the PDF manuals, ‘<span class="file">Rd.sty</span>’,
now loads the standard ‘<span class="samp">⁠amsmath⁠</span>’, ‘<span class="samp">⁠amsfonts⁠</span>’ and
‘<span class="samp">⁠amssymb⁠</span>’ packages for greater coverage of math commands in
the Rd <code style="white-space: pre;">⁠\eqn⁠</code> and <code style="white-space: pre;">⁠\deqn⁠</code> macros.
The <code style="white-space: pre;">⁠\mathscr⁠</code> LaTeX command is also provided (via the
‘<span class="samp">⁠mathrsfs⁠</span>’ package, if available, or the ‘<span class="samp">⁠amsfonts⁠</span>’
bundle otherwise), fulfilling the wish of <a href="https://bugs.R-project.org/show_bug.cgi?id=18398">PR#18398</a>.
</p>
</li>
<li><p> (Windows) The default format of <code>readClipboard()</code> and
<code>writeClipboard()</code> has been changed to <code>13</code>
(<code>CF_UNICODETEXT</code>).
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> The PDF manuals (if built) can be compacted by the new
target <code>make compact-pdf</code> (at the top level or in
directory ‘<span class="file">doc/manual</span>’).
</p>
</li>
<li><p> There is now <code>configure</code> support for LLVM clang 15
on Linux, which defaults to position-independent (PIE) executables
whereas <code>gfortran</code> does not.
</p>
</li>
<li><p> Many small changes to ease compilation (and suppress
warnings) with LLVM <code>clang</code> 15.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>Rscript -e</code> would fail if ‘<span class="file">stdin</span>’ were closed
(Reported by Henrik Bengtsson.)
</p>
</li>
<li> <p><code>qt(*, log.p=TRUE)</code> in outer tails no longer produces
<code>NaN</code> in its final steps, thus fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18360">PR#18360</a>.
</p>
</li>
<li> <p><code>tools::Rd2latex()</code> now escapes hashes and ampersands
when writing URLs, fixing LaTeX errors with such URLs in <code style="white-space: pre;">⁠\tabular⁠</code>.
</p>
</li>
<li><p> When <code>isGeneric(f, fdef=*)</code> is used with mismatching
names, the warning is better understandable; reported (with fix) in
<a href="https://bugs.R-project.org/show_bug.cgi?id=18370">PR#18370</a> by Gabe Becker.
</p>
</li>
<li> <p><code>poly(x, n)</code> now works again (and is now documented)
when <code>x</code> is a <code>"Date"</code> or <code>"POSIXct"</code> object, or of
another class while fulfilling <code>mode(x) == "numeric"</code>. This
also enables <code>poly(x, *, raw=TRUE)</code> for such variables.
Reported by Michael Chirico to R-devel.
</p>
</li>
<li> <p><code>write.table()</code>, <code>write.csv()</code> and
<code>write.csv2()</code> restore their numerical precision (internal
equivalent of <code>digits = 15</code>) after an interrupt (<a href="https://bugs.R-project.org/show_bug.cgi?id=18384">PR#18384</a>).
</p>
</li>
<li><p> One can now read also byte <code>FF</code> from a clipboard connection
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18385">PR#18385</a>).
</p>
</li>
<li> <p><code>source("")</code> and <code>source(character())</code> now give more
helpful error messages.
</p>
</li>
<li> <p><code>R CMD check --as-cran</code> set <span class="env">_R_CHECK_TIMINGS_</span>
too late to have the intended effect.
</p>
</li>
<li> <p><code>as.POSIXlt(x)</code> now also works with very large dates
<code>x</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18401">PR#18401</a> reported by Hannes Mühleisen.
</p>
</li>
<li><p> Files can now be extracted even from very large zip archives
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18390">PR#18390</a>, thanks to Martin Jakt).
</p>
</li>
<li><p> Non-finite objects of class <code>"POSIXlt"</code> are now
correctly coerced to classes <code>"Date"</code> and <code>"POSIXct"</code>;
following up on the extension to <code>format()</code> them correctly.
</p>
</li>
<li><p> Added methods for <code>is.finite()</code>, <code>is.infinite()</code>
and <code>is.nan()</code> for <code>"POSIXlt"</code> date-time objects.
</p>
</li></ul>
<h4>BUG FIXES on Windows</h4>
<ul>
<li><p> Non-ASCII characters are now properly displayed on Windows
in windows created using GraphApp via
e.g. <code>winDialogString</code> thanks to a workaround for an
at least surprising Windows behavior with UTF-8 as the system
encoding (<a href="https://bugs.R-project.org/show_bug.cgi?id=18382">PR#18382</a>).
</p>
</li>
<li><p> Find and replace operations work again in the script editor in
<code>Rgui</code> on Windows.
</p>
</li>
<li><p> Computation of window size based on requested client size in
GraphApp when running in a multi-byte locale on Windows has been fixed
(regression in <span class="rlang"><b>R</b></span> 4.2.0 for users of systems where <span class="rlang"><b>R</b></span> 4.1 used a
single-byte locale). <code>Rgui</code> again respects the number of
console rows and columns given in ‘<span class="file">Rconsole</span>’ file.
</p>
</li>
<li> <p><code>Rterm</code> support for <code>Alt+xxx</code> sequences has been
fixed to produce the corresponding character (only) once. This fixes
pasting text with tilde on Italian keyboard (<a href="https://bugs.R-project.org/show_bug.cgi?id=18391">PR#18391</a>).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.2.1</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> New function <code>utils::findCRANmirror()</code> to find out if a
<abbr>CRAN</abbr> mirror has been selected, otherwise fallback to
the main site. This behaves in the same way as
<code>tools::CRAN_package_db()</code> and is intended for packages
wishing to access <abbr>CRAN</abbr> for purposes other than
installing packages.
</p>
<p>The need for this was shown by a day when the main <abbr>CRAN</abbr>
website was offline and a dozen or so packages which had its URL
hardcoded failed their checks.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> The libraries searched for by <span class="option">--with-blas</span> (without
a value) now include BLIS (after OpenBLAS but before ATLAS). And
on macOS, the Accelerate framework (after ATLAS). (This is
patterned after the <code>AX_BLAS</code> macro from the Autoconf Archive.)
</p>
</li>
<li><p> The included LAPACK sources have been updated to 3.10.1.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li><p> The (full path to) the command <code>tidy</code> to be used for
HTML validation can be set by environment variable <span class="env">R_TIDYCMD</span>.
</p>
</li>
<li><p> Setting environment variable
<span class="env">_R_CHECK_RD_VALIDATE_RD2HTML_</span> to a false value will override
<code>R CMD check --as-cran</code> and turn off HTML validation.
This provides a way to circumvent a problematic <code>tidy</code>.
</p>
<p>The 2006 version that ships with macOS is always skipped.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The undocumented legacy declarations of <code>Sint</code>,
<code>Sfloat</code>, <code>SINT_MAX</code> and <code>SINT_MIN</code> in header
‘<span class="file">R.h</span>’ are deprecated.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>fisher.test(d)</code> no longer segfaults for “large”
<code>d</code>; fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18336">PR#18336</a> by preventing/detecting an integer
overflow reliably.
</p>
</li>
<li> <p><code>tar(., files=*)</code> now produces correctly the warning
about invalid <abbr>UID</abbr> or <abbr>GID</abbr> of files, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18344">PR#18344</a>, reported by
Martin Morgan.
</p>
</li>
<li> <p><code>tk_choose.files()</code> with <code>multi = FALSE</code>
misbehaved on paths containing spaces (<a href="https://bugs.R-project.org/show_bug.cgi?id=18334">PR#18334</a>) (regression
introduced in <span class="rlang"><b>R</b></span> 4.0.0).
</p>
</li>
<li> <p><code>sort(x, partial = ind, *)</code> now works correctly notably for
the non-default <code>na.last = FALSE</code> or <code>TRUE</code>, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18335">PR#18335</a> reported by James Edwards.
</p>
</li>
<li><p> Environment variable <span class="env">_R_CHECK_XREFS_REPOSITORIES_</span> is
only used for checking ‘<span class="file">.Rd</span>’ cross-references in <code>R
CMD check</code> (as documented) and not for other uses looking for a
<abbr>CRAN</abbr> mirror.
</p>
</li>
<li><p> The search for a <abbr>CRAN</abbr> mirror when checking
packages now uses <code>getOption("repos")</code> if that specifies a
<abbr>CRAN</abbr> mirror, even when it does not also specify all three
Bioconductor repositories (as was previously required).
</p>
</li>
<li><p> The <abbr><span class="acronym">HTML</span></abbr> code generated by <code>tools::Rd2HTML()</code>
has been improved to pass <code>tidy</code> 5.8.0.
</p>
</li></ul>
<h4>BUG FIXES on Windows</h4>
<ul>
<li><p> Writing to a clipboard connection works again, fixing a
regression in <span class="rlang"><b>R</b></span> 4.2.0 (<a href="https://bugs.R-project.org/show_bug.cgi?id=18332">PR#18332</a>). Re-using a closed clipboard
connection no longer issues a spurious warning about an ignored
encoding argument.
</p>
</li>
<li><p> C function <code>getlocale</code> no longer attempts to query an
unsupported category from the OS, even when requested at R level,
which may cause crashes when <span class="rlang"><b>R</b></span> 4.2.0 (which uses <abbr>UCRT</abbr>) is
embedded (reported by Kevin Ushey).
</p>
</li>
<li><p> Accent keys now work in GraphApp Unicode windows, which are
used by <code>Rgui</code> whenever running in a multibyte locale (so
also in UTF-8, hence fixing a regression in <span class="rlang"><b>R</b></span> 4.2.0 for users of
systems where <span class="rlang"><b>R</b></span> 4.1 used a single-byte locale).
</p>
</li>
<li><p> Completion in <code>Rgui</code> now works also with
non-ASCII characters.
</p>
</li>
<li> <p><code>Rgui</code> no longer truncates usage information with
<span class="option">--help</span>.
</p>
</li>
<li><p> Text injection from external applications via
<code>SendInput</code> now works in GraphApp Unicode windows, fixing a
regression in <span class="rlang"><b>R</b></span> 4.2.0 for <code>Rgui</code> users of systems where
<span class="rlang"><b>R</b></span> 4.1 used a single-byte locale but <span class="rlang"><b>R</b></span> 4.2.0 uses UTF-8.
</p>
</li>
<li><p> Performance of <code>txtProgressBar()</code> in <code>Rgui</code> when
running in a multi-byte locale has been improved (fixing a
performance regression in <span class="rlang"><b>R</b></span> 4.2.0 for users of systems where <span class="rlang"><b>R</b></span>
4.1 used a single-byte locale).
</p>
</li>
<li><p> The script editor in <code>Rgui</code> now works also on
systems using UTF-8 as the native encoding. Users of the script
editor have to convert their scripts with non-ASCII characters to
UTF-8 before reading them in <span class="rlang"><b>R</b></span> 4.2.1 or newer (on recent Windows
where UTF-8 is used). This fixes a regression in <span class="rlang"><b>R</b></span> 4.2.0, which
prevented some operations with scripts when they contained
non-ASCII characters.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.2.0</h3>
<h4>SIGNIFICANT USER-VISIBLE CHANGES</h4>
<ul>
<li><p> The <code>formula</code> method of <code>aggregate()</code> now matches
the generic in naming its first argument <code>x</code> (resolving
<a href="https://bugs.R-project.org/show_bug.cgi?id=18299">PR#18299</a> by Thomas Soeiro).
</p>
<p>This means that calling <code>aggregate()</code> with a formula as a
named first argument requires name <code>formula</code> in earlier
versions of <span class="rlang"><b>R</b></span> and name <code>x</code> now, so portable code should not
name the argument (code in many packages did).
</p>
</li>
<li><p> Calling <code>&&</code> or <code>||</code> with either argument of
length greater than one now gives a warning (which it is intended
will become an error).
</p>
</li>
<li><p> Calling <code>if()</code> or <code>while()</code> with a condition of
length greater than one gives an error rather than a warning.
Consequently, environment variable
<span class="env">_R_CHECK_LENGTH_1_CONDITION_</span> no longer has any effect.
</p>
</li>
<li><p> Windows users should consult the WINDOWS section below for
some profound changes including
</p>
<ul>
<li><p> Support for 32-bit builds has been dropped.
</p>
</li>
<li><p> UTF-8 locales are used where available.
</p>
</li>
<li><p> The default locations for the <span class="rlang"><b>R</b></span> installation and personal
library folder have been changed.
</p>
</li></ul>
<p>Thanks to Tomas Kalibera for months of work on the Windows port
for this release.
</p>
</li></ul>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>matrix(x, n, m)</code> now warns in more cases where
<code>length(x)</code> differs from <code>n * m</code>, as suggested by
Abby Spurdle and Wolfgang Huber in Feb 2021 on the R-devel
mailing list.
</p>
<p>This warning can be turned into an error by setting environment
variable <span class="env">_R_CHECK_MATRIX_DATA_</span> to ‘<span class="samp">⁠TRUE⁠</span>’: <code>R
CMD check --as-cran</code> does so unless it is already set.
</p>
</li>
<li><p> Function <code>file_test()</code> in package <span class="pkg">utils</span> gains
tests for symlinks, readability and writability.
</p>
</li>
<li> <p><code>capabilities("libxml")</code> is now false.
</p>
<p>The description of <code>capabilities("http/ftp")</code> now reflects
that it refers to the default method, no longer the internal one.
</p>
</li>
<li> <p><code>simplify2array()</code> gains an <code>except</code> argument for
controlling the exceptions used by <code>sapply()</code>.
</p>
</li>
<li><p> Environment variables <span class="env">R_LIBS_USER</span> and
<span class="env">R_LIBS_SITE</span> are both now set to the <span class="rlang"><b>R</b></span> system default if
unset or empty, and can be set to <code>NULL</code> to indicate an empty
list of user or site library directories.
</p>
</li>
<li><p> The warning for <code>axis()</code>(-like) calls in cases of relatively
small ranges (typically in log-scale situations) is slightly improved
<em>and</em> suppressed from explicit calls to <code>.axisPars()</code> as
has always been the intention.
</p>
</li>
<li><p> The <code>contrasts</code> setter function <code>`contrasts<-`</code>
gains an explicit default <code>how.many = NULL</code> rather than just
using <code>missing(how.many)</code>.
</p>
</li>
<li> <p><code>grid.pretty()</code> gains a new optional argument <code>n = 5</code>.
</p>
</li>
<li><p> There is a new function <code>.pretty()</code> with option
<code>bounds</code> as a technical-utility version of <code>pretty()</code>. It
and <code>pretty()</code> gain a new argument <code>f.min</code> with a better
than back-compatible default.
</p>
</li>
<li><p> Function <code>grDevices::axisTicks()</code> and related functions
such as <code>graphics::axis()</code> work better, notably for the log
scale; partly because of the <code>pretty()</code> improvements, but also
because care is taken e.g., when <code>ylim</code> is finite but
<code>diff(ylim)</code> is infinite.
</p>
</li>
<li> <p><code>nclass.FD()</code> gains a <code>digits</code> option.
</p>
</li>
<li><p> The R Mathlib internal C function <code>bd0()</code> (called
indirectly from a dozen probability density and distribution
functions such as <code>dpois()</code>, <code>dbinom()</code>,
<code>dgamma()</code>, <code>pgamma()</code> <em>etc</em>) has been complemented
by a more sophisticated and (mostly) more accurate C function
<code>ebd0()</code>, currently called only by internal
<code>dpois_raw()</code> improving accuracy for <span class="rlang"><b>R</b></span> level <code>dpois()</code>
and potentially others calling it such as <code>dnbinom()</code>,
<code>dgamma()</code> or <code>pgamma()</code>. (Thanks to Morten Welinder's
<a href="https://bugs.R-project.org/show_bug.cgi?id=15628">PR#15628</a>.)
</p>
</li>
<li> <p><code>write.ftable()</code> gains <code>sep = " "</code> argument as
suggested by Thomas Soeiro.
</p>
</li>
<li><p> The names of the locale categories supported by <span class="rlang"><b>R</b></span>'s
<code>Sys.getlocale()</code> and <code>Sys.setlocale()</code> are now provided
by variable <code>.LC.categories</code> in the <code>base</code> namespace.
</p>
</li>
<li><p> The <code>Date</code> and <code>POSIXt</code> methods for <code>hist()</code>
and the <code>histogram</code> method for <code>plot()</code> now also use the
new default <code>col = "lightgray"</code> in consistency with the
corresponding change to <code>hist()</code>'s default for <span class="rlang"><b>R</b></span> 4.0.0.
</p>
</li>
<li> <p><code>hist.default()</code> gains new <code>fuzz</code> argument, and the
histogram <code>plot</code> method no longer uses fractional axis ticks
when displaying counts (<code>"Frequency"</code>).
</p>
</li>
<li> <p><code>mapply()</code> and hence <code>Map()</code> now also obey the
“max-or-0-if-any” recycling rule, such that, e.g.,
<code>Map(`+`, 1:3, 1[0])</code> is valid now.
</p>
</li>
<li> <p><code>as.character(<obj>)</code> for <code>"hexmode"</code> or
<code>"octmode"</code> objects now fulfils the important basic rule
<code>as.character(x)[j] === as.character(x[j])</code>.
</p>
</li>
<li><p> The set utility functions, notably <code>intersect()</code> have been
tweaked to be more consistent and symmetric in their two set
arguments, also preserving a common <code>mode</code>.
</p>
</li>
<li> <p><code>substr(ch, start,end) <- new</code> now e.g., preserves
<code>names(ch)</code>; ditto for <code>substring()</code>, thanks to a patch
from Brodie Gaslam.
</p>
</li>
<li> <p><code>plot(<lm>)</code> gains a <code>extend.ylim.f</code> argument, in
partial response to <a href="https://bugs.R-project.org/show_bug.cgi?id=15285">PR#15285</a>; further <a href="https://bugs.R-project.org/show_bug.cgi?id=17784">PR#17784</a> is fixed thanks to
several contributors and a patch by Elin Waring.
The Cook's dist contours get customizable via <code>cook.col</code> and
<code>cook.lty</code> with a different default color and their legend is
nicer by default and customizable via <code>cook.legendChanges</code>.
</p>
</li>
<li><p> Attempting to subset an object that is not subsettable now
signals an error of class <code>notSubsettableError</code>. The
non-subsettable object is contained in the <code>object</code> field of
the error condition.
</p>
</li>
<li><p> Subscript-out-of-bounds errors are now signaled as errors of
class <code>subscriptOutOfBoundsError</code>.
</p>
</li>
<li><p> Stack-overflow errors are now signaled as errors inheriting
from class <code>stackOverflowError</code>. See
<code>?stackOverflowError</code> for more details.
</p>
</li>
<li><p> New partly experimental <code>Sys.setLanguage()</code> utility,
solving the main problem of <a href="https://bugs.R-project.org/show_bug.cgi?id=18055">PR#18055</a>.
</p>
</li>
<li> <p><code>gettext()</code> and <code>gettextf()</code> get a new option
<code>trim = TRUE</code> which when set to false allows translations for
strings such as <code>"Execution halted\n"</code> typical for C code.
</p>
</li>
<li><p> An experimental implementation of hash tables is now
available. See <code>?hashtab</code> for more details.
</p>
</li>
<li> <p><code>identical()</code> gains a <code>extptr.as.ref</code> argument for
requesting that external pointer objects be compared as reference
objects.
</p>
</li>
<li> <p><code>reorder()</code> gets an argument <code>decreasing</code> which it
passes to <code>sort()</code> for level creation; based on the wish and
patch by Thomas Soeiro in <a href="https://bugs.R-project.org/show_bug.cgi?id=18243">PR#18243</a>.
</p>
</li>
<li> <p><code>as.vector()</code> gains a <code>data.frame</code> method which
returns a simple named list, also clearing a long-standing
‘FIXME’ to enable <code>as.vector(<data.frame>,
mode="list")</code>. This breaks code relying on
<code>as.vector(<data.frame>)</code> to return the unchanged data frame.
</p>
</li>
<li> <p><code>legend()</code> is now vectorized for arguments <code>cex</code>,
<code>x.intersp</code>, and <code>text.width</code>.
The latter can now also be specified as a vector (one element for
each column of the legend) or as <code>NA</code> for computing a proper
column wise maximum value of <code>strwidth(legend)</code>.
The argument <code>y.intersp</code> can be specified as a vector with
one entry for each row of the legend.
</p>
<p><code>legend()</code> also gains new arguments <code>title.cex</code> and
<code>title.font</code>.
Thanks to Swetlana Herbrandt.
</p>
</li>
<li><p> Deparsing no longer remaps attribute names <code>dim</code>,
<code>dimnames</code>, <code>levels</code>, <code>names</code> and <code>tsp</code> to
historical S-compatible names (which <code>structure()</code> maps back).
</p>
</li>
<li> <p><code>sample()</code> and <code>sample.int()</code> have additional
sanity checks on their <code>size</code> and <code>n</code> arguments.
</p>
<p><code>all.equal.numeric()</code> gains a sanity check on its
<code>tolerance</code> argument – calling <code>all.equal(a, b, c)</code> for
three numeric vectors is a surprisingly common error.
</p>
<p><code>mean(na.rm =)</code>, <code>rank(na.last =)</code>,
<code>barplot(legend.text =)</code>, <code>boxplot()</code>,
<code>contour(drawlabels =)</code>, <code>polygon(border =)</code> and
<code>methods::is(class2 =)</code> have more robust sanity checks on
their arguments.
</p>
<p><code>R CMD Rd2pdf</code> (used by <code>R CMD check</code>) has a more
robust sanity check on the format of <code>\alias{}</code> commands.
</p>
</li>
<li> <p><code>psigamma(x, deriv)</code> for negative <code>x</code> now also works
for <code>deriv = 4</code> and <code>5</code>; their underlying C level
<code>dpsifn()</code> is documented in ‘Writing R Extensions’.
</p>
</li>
<li><p> The HTML help system now uses HTML5 (wish of <a href="https://bugs.R-project.org/show_bug.cgi?id=18149">PR#18149</a>).
</p>
</li>
<li> <p><code>ks.test()</code> now provides exact p-values also with ties
and MC p-values in the two-sample (Smirnov) case.
By Torsten Hothorn.
</p>
</li>
<li> <p><code>ks.test()</code> gains a formula interface, with <code>y ~ 1</code>
for the one-sample (Kolmogorov) test and <code>y ~ group</code> for the
two-sample (Smirnov) test. Contributed by Torsten Hothorn.
</p>
</li>
<li><p> The return value from <code>ks.test()</code> now has class
<code>c("ks.test", "htest")</code> – packages using <code>try()</code> need
to take care to use <code>inherits()</code> and not <code>==</code> on the class.
</p>
</li>
<li><p> New functions <code>psmirnov()</code>, <code>qsmirnov()</code> and
<code>rsmirnov()</code> in package <span class="pkg">stats</span> implementing the
asymptotic and exact distributions of the two-sample Smirnov statistic.
</p>
</li>
<li> <p><code>iconv()</code> now allows <code>sub = "c99"</code> to use C99-style
escapes for UTF-8 inputs which cannot be converted to encoding <code>to</code>.
</p>
</li>
<li><p> In a forward pipe <code>|></code> expression it is now possible to
use a named argument with the placeholder <code>_</code> in
the <code>rhs</code> call to specify where the <code>lhs</code> is to be
inserted. The placeholder can only appear once on the <code>rhs</code>.
</p>
</li>
<li><p> The included LAPACK sources have been updated to version 3.10.0,
except for the four Fortran 77 routines which 3.10.0 has
re-implemented in Fortran 90 (where the older versions have been
retained as the <span class="rlang"><b>R</b></span> build process does not support Fortran 90).
</p>
</li>
<li> <p><code>path.expand()</code> and most other uses of tilde expansion now
warn if a path would be too long if expanded. (An exception is
<code>file.exists()</code>, which silently returns false.)
</p>
</li>
<li> <p><code>trunc(<Date>, *)</code> now supports <code>units = "months"</code>
or <code>"years"</code> for consistency with the <code>POSIXt</code> method,
thanks to Dirk Eddelbuettel's proposal in <a href="https://bugs.R-project.org/show_bug.cgi?id=18099">PR#18099</a>.
</p>
</li>
<li> <p><code>list2DF()</code> now checks that its arguments are of the
same length, rather than use recycling.
</p>
</li>
<li><p> The HTML help system has several new features: LaTeX-like
math can be typeset using either <a href="https://katex.org/">KaTeX</a>
or <a href="https://www.mathjax.org/">MathJax</a>, usage and example
code is highlighted using <a href="https://prismjs.com">Prism</a>, and
for dynamic help the output of examples and demos can be shown
within the browser if the <a href="https://CRAN.R-project.org/package=knitr"><span class="pkg">knitr</span></a> package is
installed. These features can be disabled by setting the
environment variable <code>_R_HELP_ENABLE_ENHANCED_HTML_</code> to a
false value.
</p>
</li></ul>
<h4>GRAPHICS</h4>
<ul>
<li><p> The graphics engine version, <code>R_GE_version</code>, has been
bumped to <code>15</code> and so packages that provide graphics devices
should be reinstalled.
</p>
</li>
<li><p> The <span class="pkg">grid</span> package now allows the user to specify a
“vector” of pattern fills. The <code>fill</code> argument
to <code>gpar()</code> accepts a list of gradients and/or patterns and
the functions <code>linearGradient()</code>, <code>radialGradient()</code>,
and <code>pattern()</code> have a new <code>group</code> argument.
</p>
<p>Points grobs (data symbols) can now also have a pattern fill.
</p>
<p>The <code>grobCoords()</code> function now returns a more informative
and complex result.
</p>
</li>
<li><p> The <span class="pkg">grid</span> package has new functions for drawing
isolated groups: <code>grid.group()</code>, <code>grid.define()</code>, and
<code>grid.use()</code>. These functions add compositing operators and
affine transformations to R's graphics capabilities.
</p>
<p>The <span class="pkg">grid</span> package also has new functions for stroking
and filling paths: <code>grid.stroke()</code>, <code>grid.fill()</code>,
and <code>grid.fillStroke()</code>.
</p>
<p>A new function <code>as.path()</code> allows the user to specify the
fill rule for a path that is to be used for clipping, stroking, or
filling; available options are <code>"winding"</code> and
<code>"evenodd"</code>. A new function <code>as.mask()</code> allows the user
to specify the type of a mask; available options are
<code>"alpha"</code> and <code>"luminance"</code>.
</p>
<p>These new features are only supported so far (at most) on the
Cairo-based graphics devices and on the <code>pdf()</code> device.
</p>
</li>
<li> <p><code>dev.capabilities()</code> reports on device support
for the new features.
</p>
</li>
<li> <p><code>par()</code> now warns about unnamed non-character arguments
to prevent misuse such as <code>{usr <- par("usr"); par(usr)}</code>.
</p>
</li></ul>
<h4>WINDOWS</h4>
<ul>
<li> <p><span class="rlang"><b>R</b></span> uses UTF-8 as the native encoding on recent Windows
systems (at least Windows 10 version 1903, Windows Server 2022 or
Windows Server 1903). As a part of this change, R uses <abbr>UCRT</abbr> as
the C runtime. <abbr>UCRT</abbr> should be installed manually on systems older
than Windows 10 or Windows Server 2016 before installing <span class="rlang"><b>R</b></span>.
</p>
</li>
<li><p> The default personal library on Windows, folder
‘<span class="file">R\win-library\x.y</span>’ where ‘<span class="file">x.y</span>’ stands for <span class="rlang"><b>R</b></span> release
‘<span class="file">x.y.z</span>’, is now a subdirectory of Local Application Data
directory (usually a hidden directory
‘<span class="file">C:\Users\username\AppData\Local</span>’). Use
<code>shell.exec(.libPaths()[1])</code> from <span class="rlang"><b>R</b></span> to open the personal
library in Explorer when it is first in the list (<a href="https://bugs.R-project.org/show_bug.cgi?id=17842">PR#17842</a>).
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> uses a new 64-bit Tcl/Tk bundle. The previous
32-bit/64-bit bundle had a different layout and can no longer be
used.
</p>
</li>
<li><p> Make files and installer scripts for Windows have been
tailored to ‘<span class="file">Rtools42</span>’, the newly recommended 64-bit
<code>gcc</code> 10.3 MinGW-W64 <abbr>UCRT</abbr> toolchain.
</p>
</li>
<li> <p>‘<span class="file">Rtools42</span>’ by default uses the Windows security
features ASLR and DEP; hence <abbr>CRAN</abbr> builds of <span class="rlang"><b>R</b></span> and packages also
do.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> now supports files ‘<span class="file">Makevars.ucrt</span>’,
‘<span class="file">Makefile.ucrt</span>’, ‘<span class="file">configure.ucrt</span>’ and
‘<span class="file">cleanup.ucrt</span>’ in packages, which are used in preference to
the ‘<span class="file">.win</span>’ variants. This allows keeping the ‘<span class="file">.win</span>’
files around to support older versions of <span class="rlang"><b>R</b></span>. This feature will
be removed in the future once support for older versions of <span class="rlang"><b>R</b></span>
would no longer be needed.
</p>
</li>
<li> <p><code>R.version</code> gains a new field <code>crt</code> (only on
Windows) to denote the C runtime. The value is <code>"ucrt"</code>.
</p>
</li>
<li><p> On Windows, <code>download.file(method = "auto")</code> and
<code>url(method = "default")</code> now follow Unix in using
<code>"libcurl"</code> for all except ‘<span class="samp">⁠file://⁠</span>’ <abbr>URI</abbr>s.
</p>
</li>
<li> <p>‘<span class="file">Rtools42</span>’ includes an unpatched Msys2 build of GNU
<code>tar</code>. Paths including drive letters can be made to work
by adding <span class="option">--force-local</span> to environment variable
<span class="env">TAR_OPTIONS</span>. (‘<span class="file">Rtools40</span>’ and earlier included a
patched version which defaulted to this option.)
</p>
</li>
<li><p> Installer builds of <span class="rlang"><b>R</b></span> automatically find the
‘<span class="file">Rtools42</span>’ software collection as well as the compiler
toolchain. No <span class="env">PATH</span> setting is required from the user.
</p>
</li>
<li><p> The default installation directory of R for a user-only
installation has been changed to the User Program Files directory
(usually a hidden directory
‘<span class="file">C:\Users\username\AppData\Local\Programs</span>’) to follow
Windows conventions. Use <code>shell.exec(R.home())</code> from <span class="rlang"><b>R</b></span> to
open the <span class="rlang"><b>R</b></span> installation directory in Explorer (<a href="https://bugs.R-project.org/show_bug.cgi?id=17842">PR#17842</a>).
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> now supports installation-time patching of packages.
Patches may be installed from a supplied URL or a local directory
or disabled. Patches are included into the installed packages for
reference. This experimental feature may be removed in the
future.
</p>
</li>
<li> <p><code>libcurl</code> is now required for building from source.
</p>
</li>
<li><p> The clipboard connection now works also with text in other
than the current native encoding (<a href="https://bugs.R-project.org/show_bug.cgi?id=18267">PR#18267</a>, with
Hiroaki Yutani). Text is always pasted to the clipboard in
UTF16-LE and the <code>encoding</code> argument is ignored.
</p>
</li>
<li><p> The internal case-changing functions are now used by default
on Windows – this circumvents problems (for example with E acute)
of the <abbr>UCRT</abbr> Windows' runtime.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> on Windows now uses the system memory allocator. Doug
Lea's allocator was used since <span class="rlang"><b>R</b></span> 1.2.0 to mitigate performance
limitations seen with system allocators on earlier versions of
Windows.
</p>
</li>
<li> <p><code>memory.limit()</code> and <code>memory.size()</code> are now stubs on
Windows (as on Unix-alikes).
</p>
</li>
<li><p> Applications embedding <span class="rlang"><b>R</b></span> on Windows can now use additional
callbacks, which have so far only been available only on Unix
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18286">PR#18286</a>).
</p>
</li></ul>
<h4>INSTALLATION</h4>
<ul>
<li><p> Facilities for accessing ‘<span class="samp">⁠ftp://⁠</span>’ sites are no longer
tested (except <em>pro tem</em> for <code>curlGetHeaders()</code>) as modern
browsers have removed support.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> can now be built with ‘<span class="samp">⁠DEFS = -DSTRICT_R_HEADERS⁠</span>’ .
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li> <p><code>R CMD INSTALL</code> no longer tangles vignettes. This
completes an <code>R CMD build</code> change in <span class="rlang"><b>R</b></span> 3.0.0 and affects
packages built before <span class="rlang"><b>R</b></span> 3.0.2. Such packages should be re-made
with <code>R CMD build</code> to have the tangled <span class="rlang"><b>R</b></span> code of vignettes
shipped with the tarball.
</p>
</li>
<li> <p><code>USE_FC_LEN_T</code> will become the default: this uses the
correct prototypes for Fortran BLAS/LAPACK routines called from
C/C++, and requires adjustment of most such calls – see
‘Writing R Extensions’ §6.6.1. (This has been supported
since <span class="rlang"><b>R</b></span> 3.6.2.)
</p>
</li>
<li><p> Package installation speed for packages installed with
<code>keep.source</code> has been improved. This resolve the issue
reported by Ofek Shilon in <a href="https://bugs.R-project.org/show_bug.cgi?id=18236">PR#18236</a>.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> can optionally report
files/directories left behind in home, ‘<span class="file">/tmp</span>’ (even though
<span class="env">TMPDIR</span> is set) and other directories. See the “R
Internals” manual for details.
</p>
</li>
<li> <p><code>R CMD check</code> now reports byte-compilation errors
during installation. These are not usually fatal but may result
in parts of the package not being byte-compiled.
</p>
</li>
<li> <p><span class="env">_R_CHECK_DEPENDS_ONLY_</span> can be applied selectively to
examples, tests and/or vignettes in <code>R CMD check</code>: see the
“R Internals” manual.
</p>
</li>
<li> <p><span class="env">_R_CHECK_SRC_MINUS_W_IMPLICIT_</span> now defaults to true:
recent versions of Apple <code>clang</code> on macOS have made
implicit function declarations in C into a compilation error.
</p>
</li>
<li> <p><code>R CMD check --as-cran</code> makes use of the
environment variable <span class="env">AUTORECONF</span>. See the
“R Internals” manual §8 for further details.
</p>
</li>
<li> <p><code>R CMD check --use-valgrind</code> also uses
<code>valgrind</code> when re-building vignettes as some non-Sweave
vignettes unhelpfully comment out all their code when
<code>R CMD check</code> runs vignettes.
</p>
</li>
<li><p> Errors in re-building vignettes (unless there are LaTeX
errors) are reported by <code>R CMD check</code> as ‘<span class="samp">⁠ERROR⁠</span>’ rather
than ‘<span class="samp">⁠WARNING⁠</span>’ when running vignettes has been skipped (as it
frequently is in <abbr>CRAN</abbr> checks and by <span class="option">--as-cran</span>).
</p>
</li>
<li> <p><code>R CMD Rd2pdf</code> gains a <span class="option">--quiet</span> option that is
used by <code>R CMD build</code> when building the PDF package manual.
</p>
</li>
<li> <p><code>R CMD Rd2pdf</code> now always runs LaTeX in batch mode,
consistent with Texinfo <code class="reqn">\ge</code> 6.7. The <span class="option">--batch</span>
option is ignored.
</p>
</li>
<li> <p><code>R CMD build</code> and <code>R CMD check</code> now include
the Rd file name and line numbers in the error message of an
<code style="white-space: pre;">⁠\Sexpr⁠</code> evaluation failure.
</p>
</li>
<li><p> For packages using the <code style="white-space: pre;">⁠\doi⁠</code> Rd macro (now an
install-time <code style="white-space: pre;">⁠\Sexpr⁠</code>) but no other dynamic Rd content,
<code>R CMD build</code> now produces a smaller tarball and is
considerably faster – skipping temporary package installation.
</p>
</li>
<li> <p><code>R CMD check</code> can optionally (but included in
<span class="option">--as-cran</span>) validate the HTML produced from the packages
‘<span class="file">.Rd</span>’ files. See
<a href="https://blog.r-project.org/2022/04/08/enhancements-to-html-documentation/">https://blog.r-project.org/2022/04/08/enhancements-to-html-documentation/</a>:
this needs a fairly recent version of HTML Tidy to be available.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The non-API header ‘<span class="file">R_ext/R-ftp-http.h</span>’ is no longer
provided, as the entry points it covered are now all defunct.
</p>
</li>
<li><p> A number of non-API declarations and macro definitions have
been moved from the installed header ‘<span class="file">Rinternals.h</span>’ to the
internal header ‘<span class="file">Defn.h</span>’. Packages that only use entry points
and definitions documented to be part of the API as specified in
‘Writing R Extensions’ §6 should not be affected.
</p>
</li>
<li><p> The macro <code>USE_RINTERNALS</code> no longer has any effect
when compiling package code. Packages which also use
<code>R_NO_REMAP</code> will need to ensure that the remapped names are
used for calls to API functions that were formerly also made
available as macros.
</p>
</li>
<li><p> The deprecated legacy S-compatibility macros <code>PROBLEM</code>,
<code>MESSAGE</code>, <code>ERROR</code>, <code>WARN</code>, <code>WARNING</code>,
<code>RECOVER</code>, ... are no longer defined in
‘<span class="file">R_ext/RS.h</span>’ (included by ‘<span class="file">R.h</span>’). Replace these by
calls to <code>Rf_error</code> and <code>Rf_warning</code> (defined in header
‘<span class="file">R_ext/Error.h</span>’ included by ‘<span class="file">R.h</span>’).
</p>
<p>Header ‘<span class="file">R_ext/RS.h</span>’ no longer includes ‘<span class="file">R_ext/Error.h</span>’.
</p>
</li>
<li><p> Header ‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’)
when included from C++ now includes the C++ header ‘<span class="file">cfloat</span>’
rather than the C header ‘<span class="file">float.h</span>’ (now possible as C++11 is
required).
</p>
</li>
<li><p> The legacy S-compatibility macros <code>DOUBLE_*</code> in
‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’) are deprecated.
</p>
</li>
<li><p> The deprecated S-compatibility macros <code>SINGLE_*</code> in
‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’) have been
removed.
</p>
</li>
<li> <p><code>R_Calloc</code>, <code>R_Free</code> and <code>R_Realloc</code> are
preferred to their unprefixed forms and error messages now use the
prefix. These forms were introduced in <span class="rlang"><b>R</b></span> 3.4.0 and are available
even when <code>STRICT_R_HEADERS</code> is defined.
</p>
</li>
<li> <p><code>rmultinom</code> has been documented in ‘Writing R
Extensions’ §6 so is now part of the <span class="rlang"><b>R</b></span> API.
</p>
</li>
<li><p> Similarly, <code>Rtanpi</code>, called from <span class="rlang"><b>R</b></span> level <code>tanpi()</code>
is now part of the <span class="rlang"><b>R</b></span> API.
</p>
</li>
<li><p> The long-deprecated, undocumented and non-API entry point
<code>call_R</code> is no longer declared in ‘<span class="file">R_ext/RS.h</span>’ (included
by ‘<span class="file">R.h</span>’).
</p>
</li>
<li><p> The header ‘<span class="file">S.h</span>’ which has been unsupported since Jan
2016 has been removed. Use ‘<span class="file">R.h</span>’ instead.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> The (non-default and deprecated) <code>method = "internal"</code>
for <code>download.file()</code> and <code>url()</code> no longer supports
‘<span class="samp">⁠http://⁠</span>’ nor ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s. (It is used only for
‘<span class="samp">⁠file://⁠</span>’ <abbr>URI</abbr>s.)
</p>
<p>On Windows, <code>download.file(method = "wininet")</code> no longer
supports ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s. (It is no longer the default method,
which is <code>"libcurl"</code> and does.)
</p>
<p>On Windows, the deprecated <code>method = "wininet"</code> now gives a
warning for ‘<span class="samp">⁠http://⁠</span>’ and ‘<span class="samp">⁠https://⁠</span>’ <abbr>URI</abbr>s for both
<code>download.file()</code> and <code>url()</code>. (It is no longer the default
method.)
</p>
</li>
<li><p> On Windows, the command-line option <span class="option">--max-mem-size</span>
and environment variable <span class="env">R_MAX_MEM_SIZE</span> are defunct. The
memory allocation limit was important for 32-bit builds,
but these are no longer supported.
</p>
</li>
<li> <p><code>default.stringsAsFactors()</code> is now formally deprecated,
where that was only mentioned on its regular help page,
previously. So it now gives a warning if called.
</p>
</li>
<li> <p><code>unix.time()</code> is defunct now; it had been deprecated since
<span class="rlang"><b>R</b></span> 3.4.0.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Setting <code>digits = 0</code> in <code>format()</code>,
<code>print.default()</code> (and hence typically <code>print()</code>) or
<code>options()</code> is again invalid. Its behaviour was
platform-dependent, and it is unclear what “zero
significant digits” should mean (<a href="https://bugs.R-project.org/show_bug.cgi?id=18098">PR#18098</a>).
</p>
</li>
<li><p> Messages from C code in the ‘<span class="file">cairo</span>’ section of package
<span class="pkg">grDevices</span> are now also offered for translation, thanks to
Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18123">PR#18123</a>.
</p>
</li>
<li> <p><code>mean(x)</code> with finite <code>x</code> now is finite also without
"long.double" capability.
</p>
</li>
<li> <p><code>R CMD Rd2pdf</code> no longer leaves an empty build
directory behind when it aborts due to an already existing output
file. (Thanks to Sebastian Meyer's <a href="https://bugs.R-project.org/show_bug.cgi?id=18141">PR#18141</a>.)
</p>
</li>
<li> <p><code>density(x, weights = w, na.rm = TRUE)</code> when <code>anyNA(x)</code>
is true, now removes weights “in parallel” to <code>x</code>, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18151">PR#18151</a>, reported by Matthias Gondan.
Additionally, it gets a <code>subdensity</code> option.
</p>
</li>
<li><p> Conversion of <code style="white-space: pre;">⁠\Sexpr[]{<expR>}⁠</code> to LaTeX or HTML no longer
produces long blocks of empty lines when <code><expR></code> itself
contains several lines all producing empty output. Thanks to a
report and patch by Ivan Krylov posted to R-devel.
</p>
</li>
<li> <p><code>R CMD build</code> no longer fails if a package vignette
uses child documents and ‘<span class="file">inst/doc</span>’ exists. (Thanks to
Sebastian Meyer's <a href="https://bugs.R-project.org/show_bug.cgi?id=18156">PR#18156</a>.)
</p>
</li>
<li><p> When an R documentation (‘help’ source) file
‘<span class="file">man/foo.Rd</span>’ in a package has <code style="white-space: pre;">⁠\donttest{..}⁠</code> examples
with a syntax error, it is now signalled as ERROR and with correct
line numbers relating to the ‘<span class="file">*-Ex.R</span>’ file, thanks to Duncan
Murdoch and Sebastian Meyer's reports and patch proposals in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17501">PR#17501</a>.
</p>
</li>
<li><p> Improved determination of the correct translation domain in
non-base packages, addressing the combination of <a href="https://bugs.R-project.org/show_bug.cgi?id=18092">PR#18092</a> and
<a href="https://bugs.R-project.org/show_bug.cgi?id=17998">PR#17998</a> (<code>#c6</code>) with reports and <em>augmented</em> patch
#2904 by Suharto Anggono.
</p>
<p>Note that <code>"R-base"</code> is no longer the default domain e.g.,
for top-level calls to <code>gettext()</code>; rather translation needs
explicit <code>domain = *</code> specification in such cases.
</p>
</li>
<li> <p><code>identical(attrib.as.set=FALSE)</code> now works correctly with
data frames with default row names. (Thanks to Charlie Gao's
<a href="https://bugs.R-project.org/show_bug.cgi?id=18179">PR#18179</a>.)
</p>
</li>
<li> <p><code>txtProgressBar()</code> now enforces a non-zero width for
argument <code>char</code>, without which no progress can be visible.
</p>
</li>
<li> <p><code>dimnames(table(d))</code> is more consistent in the case where
<code>d</code> is a list with a single component, thanks to Thomas Soeiro's
report to R-devel.
</p>
<p>Further, <code>table(d1, d2)</code> now gives an error when <code>d1</code> and
<code>d2</code> are data frames as suggested by Thomas in <a href="https://bugs.R-project.org/show_bug.cgi?id=18224">PR#18224</a>.
</p>
</li>
<li><p> Fix for drawing semi-transparent lines and fills on the
native Windows graphics device (<a href="https://bugs.R-project.org/show_bug.cgi?id=18219">PR#18219</a> and <a href="https://bugs.R-project.org/show_bug.cgi?id=16694">PR#16694</a>).
Thanks to Nick Ray for helpful diagnosis on Bugzilla.
</p>
</li>
<li><p> The deparser now wraps sub-expressions such as <code>if(A) .. </code>
with parentheses when needed; thanks to Duncan Murdoch's <a href="https://bugs.R-project.org/show_bug.cgi?id=18232">PR#18232</a>
and Lionel Henry's patches there.
</p>
</li>
<li> <p><code>remove.packages()</code> no longer tries to uninstall
<code>Priority: base</code> packages, thanks to a report and suggestions
by Colin Fay in <a href="https://bugs.R-project.org/show_bug.cgi?id=18227">PR#18227</a>.
</p>
</li>
<li> <p><code>win.metafile()</code> now has <code>xpinch</code> and
<code>ypinch</code> arguments so that the user can override Windows'
(potentially wrong) guess at device dimensions.
</p>
</li>
<li> <p><code>x[i]</code> and <code>x[[i]]</code> for non-integer <code>i</code>
should now behave in all cases as always documented: the index used is
equivalent to <code>as.integer(i)</code> unless that would overflow where
<code>trunc(i)</code> is used instead; thanks to Suharto Anggono's report
and patch proposals in <a href="https://bugs.R-project.org/show_bug.cgi?id=17977">PR#17977</a>.
</p>
</li>
<li> <p><code>asOneSidedFormula()</code> now associates the resulting
formula with the global environment rather than the evaluation
environment created for the call.
</p>
</li>
<li> <p><code><bibentry>$name</code> now matches the field name
case-insensitively, consistent with <code>bibentry()</code> creation and
the replacement method.
</p>
</li>
<li> <p><code>cbind()</code> failed to detect some length mismatches with
a mixture of time-series and non-time-series inputs.
</p>
</li>
<li><p> The default LaTeX style file ‘<span class="file">Sweave.sty</span>’ used by the
<code>RweaveLatex</code> driver no longer loads the obsolete ‘<span class="samp">⁠ae⁠</span>’
package; thanks to a report by Thomas Soeiro in <a href="https://bugs.R-project.org/show_bug.cgi?id=18271">PR#18271</a>.
Furthermore, it now skips ‘<span class="samp">⁠\usepackage[T1]{fontenc}⁠</span>’ for
engines other than pdfTeX (if detected) or if the new
‘<span class="samp">⁠[nofontenc]⁠</span>’ option is used.
</p>
</li>
<li> <p><code>smooth.spline()</code> now stores its logical <code>cv</code>
argument more safely, fixing a rare bug when printing, and also
stores <code>n</code>.
</p>
</li>
<li> <p><code>smooth.spline(x,y,*)</code> now computes the <code>cv.crit</code>
statistic correctly, also when <code>is.unsorted(x)</code>, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18294">PR#18294</a>.
</p>
</li>
<li><p> The <code>data.frame</code> method of <code>rbind()</code> now warns
when binding not-wholly-recycling vectors, by analogy to the default
method (for matrices).
</p>
</li>
<li> <p><code>setAs()</code> finds the correct class for name <code>to</code>
when multiple packages define a class with that name. Thanks to
Gabor Csardi for the report.
</p>
</li>
<li><p> Fix for detaching a package when two classes of the same name
are present in method signatures for the same generic. Thanks to
Gabor Csardi for the report.
</p>
</li>
<li> <p><code>match.arg("", c("", "a", "B"))</code> gives a better error
message, in part from <a href="https://bugs.R-project.org/show_bug.cgi?id=17959">PR#17959</a>, thanks to Elin Waring.
</p>
</li>
<li> <p><code>R CMD Sweave --clean</code> no longer removes
pre-existing files or subdirectories (<a href="https://bugs.R-project.org/show_bug.cgi?id=18242">PR#18242</a>).
</p>
</li>
<li><p> The <code>quartz()</code> device no longer splits polylines into
subpaths. That has caused narrowly-spaced lines with many points
to always look solid even when dashed line type was used due to
dash phase restarts.
</p>
</li>
<li><p> Deparsing constructs such as <code>quote(1 + `!`(2) + 3)</code> works
again as before R 3.5.0, thanks to the report and patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18284">PR#18284</a>
by Suharto Anggono.
</p>
</li>
<li> <p><code>as.list(f)</code> for a <code>factor</code> <code>f</code> now keeps
<code>names(f)</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18309">PR#18309</a>.
</p>
</li>
<li> <p><code>qbeta(.001, .9, .009)</code> and analogous <code>qf()</code> calls now
return a correct value instead of <code>NaN</code> or wrongly <code>1</code>, all
with a warning; thanks to the report by Ludger Goeminne in <a href="https://bugs.R-project.org/show_bug.cgi?id=18302">PR#18302</a>.
</p>
</li>
<li> <p><code>plot.lm()</code> failed to produce the plot of residuals vs.
factor levels (i.e., <code>which=5</code> when leverages are constant)
for models with character predictors (<a href="https://bugs.R-project.org/show_bug.cgi?id=17840">PR#17840</a>).
</p>
</li>
<li> <p><code>interaction.plot(..., xtick = TRUE)</code> misplaced the
x-axis line (<a href="https://bugs.R-project.org/show_bug.cgi?id=18305">PR#18305</a>).
</p>
</li>
<li><p> Not strictly fixing a bug, <code>format()</code>ing and
<code>print()</code>ing of non-finite <code>Date</code> and <code>POSIXt</code>
values <code>NaN</code> and <code class="reqn">\pm</code><code>Inf</code> no longer show as
<code>NA</code> but the respective string, e.g., <code>Inf</code>, for
consistency with numeric vector's behaviour, fulfilling the wish
of <a href="https://bugs.R-project.org/show_bug.cgi?id=18308">PR#18308</a>.
</p>
</li>
<li> <p><code>R CMD check</code> no longer runs test scripts generated
from corresponding ‘<span class="file">.Rin</span>’ files twice and now signals an
ERROR if processing an ‘<span class="file">.Rin</span>’ script fails.
</p>
</li>
<li> <p><code>tools::Rd2txt()</code> used for plain-text help pages now renders
<code style="white-space: pre;">⁠\href⁠</code>s (if <code>tools::Rd2txt_options(showURLs = TRUE)</code>)
and <code style="white-space: pre;">⁠\url⁠</code>s with percent-encoding and standards-compliant
delimiting style (angle brackets and no ‘<span class="samp">⁠URL: ⁠</span>’ prefix).
<code style="white-space: pre;">⁠\email⁠</code> is now rendered with a ‘<span class="samp">⁠mailto:⁠</span>’ prefix.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.1.3</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> The default version of Bioconductor has been changed to
3.14. (This is used by <code>setRepositories</code> and the menus in
GUIs.)
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check --as-cran</code> has a workaround for a bug
in versions of <code>file</code> up to at least 5.41 which
mis-identify <abbr>DBF</abbr> files last changed in 2022 as executables.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The legacy S-compatibility macros <code>SINGLE_*</code> in
‘<span class="file">R_ext/Constants.h</span>’ (included by ‘<span class="file">R.h</span>’) are deprecated
and will be removed in <span class="rlang"><b>R</b></span> 4.2.0.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Initialization of self-starting <code>nls()</code> models with
initialization functions following the pre-R-4.1.0 API
(without the <code>...</code> argument) works again for now, with a
deprecation warning.
</p>
</li>
<li><p> Fixed quoting of <code>~autodetect~</code> in Java setting defaults
to avoid inadvertent user lookup due to leading <code>~</code>, reported
in <a href="https://bugs.R-project.org/show_bug.cgi?id=18231">PR#18231</a> by Harold Gutch.
</p>
</li>
<li> <p><code>substr(., start, stop) <- v</code> now treats <em>negative</em>
<code>stop</code> values correctly. Reported with a patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18228">PR#18228</a> by
Brodie Gaslam.
</p>
</li>
<li><p> Subscripting an array <code>x</code> without dimnames by a
<code>length(dim(x))</code>-column character matrix gave "random"
non-sense, now an error; reported in <a href="https://bugs.R-project.org/show_bug.cgi?id=18244">PR#18244</a> by Mikael Jagan.
</p>
</li>
<li> <p><code>...names()</code> now matches <code>names(list(...))</code> closely,
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18247">PR#18247</a>.
</p>
</li>
<li> <p><code>all.equal(*, scale = s)</code> now works as intended when
<code>length(s) > 1</code>, partly thanks to Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18272">PR#18272</a>.
</p>
</li>
<li> <p><code>print(x)</code> for long vectors <code>x</code> now also works for
named atomic vectors or lists and prints the correct number when
reaching the <code>getOption("max.print")</code> limit; partly thanks to a
report and proposal by Hugh Parsonage to the R-devel list.
</p>
</li>
<li> <p><code>all.equal(<selfStart>, *)</code> no longer signals a deprecation
warning.
</p>
</li>
<li> <p><code>reformulate(*, response=r)</code> gives a helpful error message
now when <code>length(r) > 1</code>, thanks to Bill Dunlap's <a href="https://bugs.R-project.org/show_bug.cgi?id=18281">PR#18281</a>.
</p>
</li>
<li><p> Modifying <code>globalCallingHandlers</code> inside
<code>withCallingHandlers()</code> now works or fails correctly, thanks to
Henrik Bengtsson's <a href="https://bugs.R-project.org/show_bug.cgi?id=18257">PR#18257</a>.
</p>
</li>
<li> <p><code>hist(<Date>, breaks = "days")</code> and
<code>hist(<POSIXt>, breaks = "secs")</code>
no longer fail for inputs of length 1.
</p>
</li>
<li> <p><code>qbeta(.001, .9, .009)</code> and similar cases now converge
correctly thanks to Ben Bolker's report in <a href="https://bugs.R-project.org/show_bug.cgi?id=17746">PR#17746</a>.
</p>
</li>
<li> <p><code>window(x, start, end)</code> no longer wrongly signals
“'start' cannot be after 'end'”, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17527">PR#17527</a> and
<a href="https://bugs.R-project.org/show_bug.cgi?id=18291">PR#18291</a>.
</p>
</li>
<li> <p><code>data()</code> now checks that its (rarely used) <code>list</code>
argument is a character vector – a couple of packages passed other
types and gave incorrect results.
</p>
</li>
<li> <p><code>which()</code> now checks its <code>arr.ind</code> argument is
TRUE rather coercing to logical and taking the first element –
which gave incorrect results in package code.
</p>
</li>
<li> <p><code>model.weights()</code> and <code>model.offset()</code> more carefully
extract their model components, thanks to Ben Bolker and Tim Taylor's
R-devel post.
</p>
</li>
<li> <p><code>list.files(recursive = TRUE)</code> now shows all broken
symlinks (previously, some of them may have been omitted, <a href="https://bugs.R-project.org/show_bug.cgi?id=18296">PR#18296</a>).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.1.2</h3>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The workaround in headers ‘<span class="file">R.h</span>’ and ‘<span class="file">Rmath.h</span>’
(<code>using namespace std;</code>) for the Oracle Developer Studio
compiler is no longer needed now C++11 is required so has been
removed. A couple more usages of <code>log()</code> (which should have
been <code>std::log()</code>) with an <code>int</code> argument are reported on
Solaris.
</p>
</li>
<li><p> The undocumented limit of 4095 bytes on messages from the
S-compatibility macros <code>PROBLEM</code> and <code>MESSAGE</code> is now
documented and longer messages will be silently truncated rather
than potentially causing segfaults.
</p>
</li>
<li><p> If the <code>R_NO_SEGV_HANDLER</code> environment variable is
non-empty, the signal handler for SEGV/ILL/BUS signals (which
offers recovery user interface) is not set. This allows more
reliable debugging of crashes that involve the console.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> The legacy S-compatibility macros <code>PROBLEM</code>,
<code>MESSAGE</code>, <code>ERROR</code>, <code>WARN</code>, <code>WARNING</code>,
<code>RECOVER</code>, ... are deprecated and will be hidden in <span class="rlang"><b>R</b></span>
4.2.0. <span class="rlang"><b>R</b></span>'s native interface of <code>Rf_error</code> and
<code>Rf_warning</code> has long been preferred.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>.mapply(F, dots, .)</code> no longer segfaults when
<code>dots</code> is not a <code>list</code> and uses <code>match.fun(F)</code> as
always documented; reported by Andrew Simmons in <a href="https://bugs.R-project.org/show_bug.cgi?id=18164">PR#18164</a>.
</p>
</li>
<li> <p><code>hist(<Date>, ...)</code> and <code>hist(<POSIXt>, ...)</code>
no longer pass arguments for <code>rect()</code> (such as <code>col</code> and
<code>density</code>) to <code>axis()</code>. (Thanks to Sebastian Meyer's
<a href="https://bugs.R-project.org/show_bug.cgi?id=18171">PR#18171</a>.)
</p>
</li>
<li> <p><code style="white-space: pre;">⁠\Sexpr{ch}⁠</code> now preserves <code>Encoding(ch)</code>. (Thanks to
report and patch by Jeroen Ooms in <a href="https://bugs.R-project.org/show_bug.cgi?id=18152">PR#18152</a>.)
</p>
</li>
<li><p> Setting the RNG to <code>"Marsaglia-Multicarry"</code> e.g., by
<code>RNGkind()</code>, now warns in more places, thanks to
André Gillibert's report and patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18168">PR#18168</a>.
</p>
</li>
<li> <p><code>gray(numeric(), alpha=1/2)</code> no longer segfaults, fixing
<a href="https://bugs.R-project.org/show_bug.cgi?id=18183">PR#18183</a>, reported by Till Krenz.
</p>
</li>
<li><p> Fixed <code>dnbinom(x, size=<very_small>, .., log=TRUE)</code>
regression, reported by Martin Morgan.
</p>
</li>
<li> <p><code>as.Date.POSIXlt(x)</code> now keeps <code>names(x)</code>, thanks to
Davis Vaughan's report and patch in <a href="https://bugs.R-project.org/show_bug.cgi?id=18188">PR#18188</a>.
</p>
</li>
<li> <p><code>model.response()</code> now strips an <code>"AsIs"</code> class typically,
thanks to Duncan Murdoch's report and other discussants in <a href="https://bugs.R-project.org/show_bug.cgi?id=18190">PR#18190</a>.
</p>
</li>
<li> <p><code>try()</code> is considerably faster in case of an error and
long call, as e.g., from some <code>do.call()</code>. Thanks to
Alexander Kaever's suggestion posted to R-devel.
</p>
</li>
<li> <p><code>qqline(y = <object>)</code> such as <code>y=I(.)</code>, now works,
see also <a href="https://bugs.R-project.org/show_bug.cgi?id=18190">PR#18190</a>.
</p>
</li>
<li><p> Non-integer <code>mgp</code> <code>par()</code> settings are now handled
correctly in <code>axis()</code> and <code>mtext()</code>, thanks to Mikael
Jagan and Duncan Murdoch's report and suggestion in <a href="https://bugs.R-project.org/show_bug.cgi?id=18194">PR#18194</a>.
</p>
</li>
<li> <p><code>formatC(x)</code> returns length zero <code>character()</code> now,
rather than <code>""</code> when <code>x</code> is of length zero, as documented,
thanks to Davis Vaughan's post to R-devel.
</p>
</li>
<li> <p><code>removeSource(fn)</code> now retains (other) <code>attributes(fn)</code>.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.1.1</h3>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>require(<var>pkg</var>, quietly = TRUE)</code> is quieter and in
particular does not warn if the package is not found.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> Use of ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s should be regarded as
deprecated, with on-going support confined to <code>method =
"libcurl"</code> and not routinely tested. (Nowadays no major browser
supports them.)
</p>
</li>
<li><p> The non-default <code>method = "internal"</code> is deprecated for
‘<span class="samp">⁠http://⁠</span>’ and ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s for both
<code>download.file</code> and <code>url</code>.
</p>
</li>
<li><p> On Windows, <code>method = "wininet"</code> is deprecated for
‘<span class="samp">⁠http://⁠</span>’, ‘<span class="samp">⁠https://⁠</span>’ and ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s for both
<code>download.file</code> and <code>url</code>. (A warning is only given for
‘<span class="samp">⁠ftp://⁠</span>’.)
</p>
<p>For ‘<span class="samp">⁠ftp://⁠</span>’ <abbr>URI</abbr>s the default method is now <code>"libcurl"</code>
if available (which it is on <abbr>CRAN</abbr> builds).
</p>
<p><code>method = "wininet"</code> remains the default for ‘<span class="samp">⁠http://⁠</span>’
and ‘<span class="samp">⁠https://⁠</span>’ <abbr>URI</abbr>s but if <code>libcurl</code> is available, using
<code>method = "libcurl"</code> is preferred.
</p>
</li></ul>
<h4>INSTALLATION</h4>
<ul>
<li> <p><code>make check</code> now works also without a LaTeX
installation. (Thanks to Sebastian Meyer's <a href="https://bugs.R-project.org/show_bug.cgi?id=18103">PR#18103</a>.)
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>make check-devel</code> works again in an <span class="rlang"><b>R</b></span> build
configured with <span class="option">--without-recommended-packages</span>.
</p>
</li>
<li> <p><code>qnbinom(p, size, mu)</code> for large <code>size/mu</code> is correct
now in a range of cases (<a href="https://bugs.R-project.org/show_bug.cgi?id=18095">PR#18095</a>); similarly for the <code>(size,
prob)</code> parametrization of the negative binomial. Also <code>qpois()</code>
and <code>qbinom()</code> are better and or faster for extreme cases. The
underlying C code has been modularized and is common to all four cases
of discrete distributions.
</p>
</li>
<li> <p><code>gap.axis</code> is now part of the <code>axis()</code> arguments which
are passed from <code>bxp()</code>, and hence <code>boxplot()</code>. (Thanks to
Martin Smith's report and suggestions in <a href="https://bugs.R-project.org/show_bug.cgi?id=18109">PR#18109</a>.)
</p>
</li>
<li> <p><code>.First</code> and <code>.Last</code> can again be set from the site
profile.
</p>
</li>
<li> <p><code>seq.int(from, to, *)</code> and <code>seq.default(..)</code> now work
better in large range cases where <code>from-to</code> is infinite where
the two boundaries are finite.
</p>
</li>
<li> <p><code>all.equal(x,y)</code> now returns <code>TRUE</code> correctly also
when several entries of <code>abs(x)</code> and <code>abs(y)</code> are close to
<code>.Machine$double.xmax</code>, the largest finite <code>numeric</code>.
</p>
</li>
<li> <p><code>model.frame()</code> now clears the object bit when removing the
<code>class</code> attribute of a value via <code>na.action</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18100">PR#18100</a>).
</p>
</li>
<li> <p><code>charClass()</code> now works with multi-character
strings on Windows (<a href="https://bugs.R-project.org/show_bug.cgi?id=18104">PR#18104</a>, fixed by Bill Dunlap).
</p>
</li>
<li> <p><code>encodeString()</code> on Solaris now works again in Latin-1
encoding on characters represented differently in UTF-8. Support for
surrogate pairs on Solaris has been improved.
</p>
</li>
<li> <p><code>file.show()</code> on Windows now works with non-ASCII path
names representable in the current native encoding (<a href="https://bugs.R-project.org/show_bug.cgi?id=18132">PR#18132</a>).
</p>
</li>
<li><p> Embedded <span class="rlang"><b>R</b></span> on Windows can now find <span class="rlang"><b>R</b></span> home directory via the
registry even when installed only for the current user (<a href="https://bugs.R-project.org/show_bug.cgi?id=18135">PR#18135</a>).
</p>
</li>
<li> <p><code>pretty(x)</code> with finite <code>x</code> now returns finite values
also in the case where the extreme <code>x</code> values are close in size
to the maximal representable number <code>.Machine$double.xmax</code>.
</p>
<p>Also, it's been tweaked for very small ranges and when a boundary is
close (or equal) to zero; e.g., <code>pretty(c(0,1e-317))</code> no longer
has negative numbers, currently still warning about a very small
range, and <code>pretty(2^-(1024 - 2^-1/(c(24,10))))</code> is more accurate.
</p>
</li>
<li><p> The error message for not finding vignette files when weaving
has correct file sizes now. (Thanks to Sebastian Meyer's <a href="https://bugs.R-project.org/show_bug.cgi?id=18154">PR#18154</a>.)
</p>
</li>
<li> <p><code>dnbinom(20, <large>, 1)</code> now correctly gives 0, and
similar cases are more accurate with underflow precaution.
(Reported by Francisco Vera Alcivar in <a href="https://bugs.R-project.org/show_bug.cgi?id=18072">PR#18072</a>.)
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.1.0</h3>
<h4>FUTURE DIRECTIONS</h4>
<ul>
<li><p> It is planned that the 4.1.x series will be the last to
support 32-bit Windows, with production of binary packages for
that series continuing until early 2023.
</p>
</li></ul>
<h4>SIGNIFICANT USER-VISIBLE CHANGES</h4>
<ul>
<li><p> Data set <code>esoph</code> in package <span class="pkg">datasets</span> now provides
the correct numbers of controls; previously it had the numbers of
cases added to these. (Reported by Alexander Fowler in <a href="https://bugs.R-project.org/show_bug.cgi?id=17964">PR#17964</a>.)
</p>
</li></ul>
<h4>NEW FEATURES</h4>
<ul>
<li> <p>‘<span class="samp">⁠www.omegahat.net⁠</span>’ is no longer one of the repositories
known by default to <code>setRepositories()</code>. (Nowadays it only
provides source packages and is often unavailable.)
</p>
</li>
<li><p> Function <code>package_dependencies()</code> (in package
<span class="pkg">tools</span>) can now use different dependency types for direct and
recursive dependencies.
</p>
</li>
<li><p> The checking of the size of tarball in
<code>R CMD check --as-cran <pkg></code> may be tweaked via the new
environment variable
<span class="env">_R_CHECK_CRAN_INCOMING_TARBALL_THRESHOLD_</span>, as suggested in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17777">PR#17777</a> by Jan Gorecki.
</p>
</li>
<li><p> Using <code>c()</code> to combine a factor with other factors now
gives a factor, an ordered factor when combining ordered factors
with identical levels.
</p>
</li>
<li> <p><code>apply()</code> gains a <code>simplify</code> argument to allow
disabling of simplification of results.
</p>
</li>
<li><p> The <code>format()</code> method for class <code>"ftable"</code> gets a
new option <code>justify</code>. (Suggested by Thomas Soeiro.)
</p>
</li>
<li><p> New <code>...names()</code> utility. (Proposed by Neal Fultz in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17705">PR#17705</a>.)
</p>
</li>
<li> <p><code>type.convert()</code> now warns when its <code>as.is</code> argument
is not specified, as the help file always said it <em>should</em>. In
that case, the default is changed to <code>TRUE</code> in line with its
change in <code>read.table()</code> (related to <code>stringsAsFactors</code>) in
<span class="rlang"><b>R</b></span> 4.0.0.
</p>
</li>
<li><p> When printing list arrays, classed objects are now shown
<em>via</em> their <code>format()</code> value if this is a short enough
character string, or by giving the first elements of their class
vector and their length.
</p>
</li>
<li> <p><code>capabilities()</code> gets new entry <code>"Rprof"</code> which is
<code>TRUE</code> when <span class="rlang"><b>R</b></span> has been configured with the equivalent of
<code>--enable-R-profiling</code> (as it is by default). (Related to
Michael Orlitzky's report <a href="https://bugs.R-project.org/show_bug.cgi?id=17836">PR#17836</a>.)
</p>
</li>
<li> <p><code>str(xS4)</code> now also shows extraneous attributes of an
S4 object <code>xS4</code>.
</p>
</li>
<li><p> Rudimentary support for vi-style tags in <code>rtags()</code> and
<code>R CMD rtags</code> has been added. (Based on a patch from
Neal Fultz in <a href="https://bugs.R-project.org/show_bug.cgi?id=17214">PR#17214</a>.)
</p>
</li>
<li> <p><code>checkRdContents()</code> is now exported from <span class="pkg">tools</span>; it
and also <code>checkDocFiles()</code> have a new option <code>chkInternal</code>
allowing to check Rd files marked with keyword <code>"internal"</code> as
well. The latter can be activated for <code>R CMD check</code> via
environment variable <span class="env">_R_CHECK_RD_INTERNAL_TOO_</span>.
</p>
</li>
<li><p> New functions <code>numToBits()</code> and <code>numToInts()</code>
extend the <code>raw</code> conversion utilities to (double precision)
<code>numeric</code>.
</p>
</li>
<li><p> Functions <code>URLencode()</code> and <code>URLdecode()</code> in
package <span class="pkg">utils</span> now work on vectors of <abbr>URI</abbr>s.
(Based on patch from Bob Rudis submitted with <a href="https://bugs.R-project.org/show_bug.cgi?id=17873">PR#17873</a>.)
</p>
</li>
<li> <p><code>path.expand()</code> can expand ‘<span class="samp">⁠~user⁠</span>’ on most
Unix-alikes even when <code>readline</code> is not in use. It tries
harder to expand ‘<span class="samp">⁠~⁠</span>’, for example should environment variable
<span class="env">HOME</span> be unset.
</p>
</li>
<li><p> For HTML help (both dynamic and static), Rd file links to
help pages in external packages are now treated as references to
topics rather than file names, and fall back to a file link only
if the topic is not found in the target package. The earlier rule
which prioritized file names over topics can be restored by
setting the environment variable <span class="env">_R_HELP_LINKS_TO_TOPICS_</span> to
a false value.
</p>
</li>
<li> <p><code>c()</code> now removes <code>NULL</code> arguments before
dispatching to methods, thus simplifying the implementation of
<code>c()</code> methods, <em>but</em> for back compatibility keeps
<code>NULL</code> when it is the first argument. (From a report and
patch proposal by Lionel Henry in <a href="https://bugs.R-project.org/show_bug.cgi?id=17900">PR#17900</a>.)
</p>
</li>
<li> <p><code>Vectorize()</code>'s result function's environment no longer
keeps unneeded objects.
</p>
</li>
<li><p> Function <code>...elt()</code> now propagates visibility
consistently with <code>..n</code>. (Thanks to Lionel Henry's
<a href="https://bugs.R-project.org/show_bug.cgi?id=17905">PR#17905</a>.)
</p>
</li>
<li> <p><code>capture.output()</code> no longer uses non-standard
evaluation to evaluate its arguments. This makes evaluation of
functions like <code>parent.frame()</code> more consistent. (Thanks to
Lionel Henry's <a href="https://bugs.R-project.org/show_bug.cgi?id=17907">PR#17907</a>.)
</p>
</li>
<li> <p><code>packBits(bits, type="double")</code> now works as inverse of
<code>numToBits()</code>. (Thanks to Bill Dunlap's proposal in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17914">PR#17914</a>.)
</p>
</li>
<li> <p><code>curlGetHeaders()</code> has two new arguments,
<code>timeout</code> to specify the timeout for that call (overriding
<code>getOption("timeout")</code>) and <code>TLS</code> to specify the minimum
<abbr>TLS</abbr> protocol version to be used for <code>https://</code> <abbr>URI</abbr>s
(<em>inter alia</em> providing a means to check for sites using
deprecated <abbr>TLS</abbr> versions 1.0 and 1.1).
</p>
</li>
<li><p> For <code>nls()</code>, an optional constant <code>scaleOffset</code>
may be added to the denominator of the relative offset convergence
test for cases where the fit of a model is expected to be exact,
thanks to a proposal by John Nash. <code>nls(*, trace=TRUE)</code> now
also shows the convergence criterion.
</p>
</li>
<li><p> Numeric differentiation <em>via</em> <code>numericDeriv()</code>
gets new optional arguments <code>eps</code> and <code>central</code>, the
latter for taking central divided differences. The latter can be
activated for <code>nls()</code> via <code>nls.control(nDcentral =
TRUE)</code>.
</p>
</li>
<li> <p><code>nls()</code> now passes the <code>trace</code> and <code>control</code>
arguments to <code>getInitial()</code>, notably for all self-starting models,
so these can also be fit in zero-noise situations via a
<code>scaleOffset</code>. For this reason, the <code>initial</code> function of a
<code>selfStart</code> model must now have <code>...</code> in its argument list.
</p>
</li>
<li> <p><code>bquote(splice = TRUE)</code> can now splice expression
vectors with attributes: this makes it possible to splice the
result of <code>parse(keep.source = TRUE)</code>. (Report and patch
provided by Lionel Henry in <a href="https://bugs.R-project.org/show_bug.cgi?id=17869">PR#17869</a>.)
</p>
</li>
<li> <p><code>textConnection()</code> gets an optional <code>name</code> argument.
</p>
</li>
<li> <p><code>get()</code>, <code>exists()</code>, and <code>get0()</code> now signal
an error if the first argument has length greater than 1.
Previously additional elements were silently ignored. (Suggested
by Antoine Fabri on R-devel.)
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> now provides a shorthand notation for creating functions,
e.g. <code>\(x) x + 1</code> is parsed as <code>function(x) x + 1</code>.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> now provides a simple native forward pipe syntax
<code>|></code>. The simple form of the forward pipe inserts the
left-hand side as the first argument in the right-hand side call.
The pipe implementation as a syntax transformation was motivated
by suggestions from Jim Hester and Lionel Henry.
</p>
</li>
<li> <p><code>all.equal(f, g)</code> for <code>function</code>s now by default also
compares their <code>environment(.)</code>s, notably via new
<code>all.equal</code> method for class <code>function</code>. Comparison of
<code>nls()</code> fits, e.g., may now need <code>all.equal(m1, m2,
check.environment = FALSE)</code>.
</p>
</li>
<li> <p><code>.libPaths()</code> gets a new option <code>include.site</code>,
allowing to <em>not</em> include the site library. (Thanks to
Dario Strbenac's suggestion and Gabe Becker's <a href="https://bugs.R-project.org/show_bug.cgi?id=18016">PR#18016</a>.)
</p>
</li>
<li><p> Lithuanian translations are now available. (Thanks to
Rimantas Žakauskas.)
</p>
</li>
<li> <p><code>names()</code> now works for <code>DOTSXP</code> objects. On the
other hand, in ‘<span class="file">R-lang</span>’, the R language manual, we now warn
against relying on the structure or even existence of such
dot-dot-dot objects.
</p>
</li>
<li> <p><code>all.equal()</code> no longer gives an error on <code>DOTSXP</code>
objects.
</p>
</li>
<li> <p><code>capabilities("cairo")</code> now applies only to the
file-based devices as it is now possible (if very unusual) to
build <span class="rlang"><b>R</b></span> with Cairo support for those but not for <code>X11()</code>.
</p>
</li>
<li><p> There is optional support for tracing the progress of
<code>loadNamespace()</code> — see its help.
</p>
</li>
<li><p> (Not Windows.)
<code>l10n_info()</code> reports an additional element, the name of the
encoding as reported by the OS (which may differ from the
encoding part (if any) of the result from
<code>Sys.getlocale("LC_CTYPE")</code>).
</p>
</li>
<li><p> New function <code>gregexec()</code> which generalizes <code>regexec()</code>
to find <em>all</em> disjoint matches and all substrings
corresponding to parenthesized subexpressions of the given regular
expression. (Contributed by Brodie Gaslam.)
</p>
</li>
<li><p> New function <code>charClass()</code> in package <span class="pkg">utils</span> to
query the wide-character classification functions in use (such as
<code>iswprint</code>).
</p>
</li>
<li><p> The names of <code>quantile()</code>'s result no longer depend on the
global <code>getOption("digits")</code>, but <code>quantile()</code> gets a new
optional argument <code>digits = 7</code> instead.
</p>
</li>
<li> <p><code>grep()</code>, <code>sub()</code>, <code>regexp</code> and variants work
considerably faster for long factors with few levels. (Thanks to
Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=18063">PR#18063</a>.)
</p>
</li>
<li><p> Provide grouping of <code>x11()</code> graphics windows within
a window manager such as <code>Gnome</code> or <code>Unity</code>; thanks to a
patch by Ivan Krylov posted to R-devel.
</p>
</li>
<li><p> The <code>split()</code> method for class <code>data.frame</code> now
allows the <code>f</code> argument to be specified as a formula.
</p>
</li>
<li> <p><code>sprintf</code> now warns on arguments unused by the format
string.
</p>
</li>
<li><p> New palettes <code>"Rocket"</code> and <code>"Mako"</code> for
<code>hcl.colors()</code> (approximating palettes of the same name
from the <span class="pkg">viridisLite</span> package).
</p>
<p>Contributed by Achim Zeileis.
</p>
</li>
<li><p> The base environment and its namespace are now locked (so
one can no longer add bindings to these or remove from these).
</p>
</li>
<li> <p><code>Rterm</code> handling of multi-byte characters has been
improved, allowing use of such characters when supported by the
current locale.
</p>
</li>
<li> <p><code>Rterm</code> now accepts <code>ALT+ +xxxxxxxx</code> sequences to
enter Unicode characters as hex digits.
</p>
</li>
<li><p> Environment variable <span class="env">LC_ALL</span> on Windows now takes
precedence over <span class="env">LC_CTYPE</span> and variables for other supported
categories, matching the POSIX behaviour.
</p>
</li>
<li> <p><code>duplicated()</code> and <code>anyDuplicated()</code> are now
optimized for integer and real vectors that are known to be sorted
via the ALTREP framework. Contributed by Gabriel Becker via <a href="https://bugs.R-project.org/show_bug.cgi?id=17993">PR#17993</a>.
</p>
</li></ul>
<h4>GRAPHICS</h4>
<ul>
<li><p> The graphics engine version, <code>R_GE_version</code>, has been
bumped to <code>14</code> and so packages that provide graphics devices
should be reinstalled.
</p>
</li>
<li><p> Graphics devices should now specify <code>deviceVersion</code> to
indicate what version of the graphics engine they support.
</p>
</li>
<li><p> Graphics devices can now specify <code>deviceClip</code>. If
<code>TRUE</code>, the graphics engine will never perform any clipping
of output itself.
</p>
<p>The clipping that the graphics engine does perform (for both
<code>canClip = TRUE</code> and <code>canClip = FALSE</code>) has been
improved to avoid producing unnecessary artifacts in clipped
output.
</p>
</li>
<li><p> The <span class="pkg">grid</span> package now allows <code>gpar(fill)</code> to be
a <code>linearGradient()</code>, a <code>radialGradient()</code>, or a
<code>pattern()</code>. The <code>viewport(clip)</code> can now also be
a grob, which defines a clipping path, and there is a new
<code>viewport(mask)</code> that can also be a grob, which defines
a mask.
</p>
<p>These new features are only supported so far on the Cairo-based
graphics devices and on the <code>pdf()</code> device.
</p>
</li>
<li><p> (Not Windows.)
A warning is given when a Cairo-based type is specified for a
<code>png()</code>, <code>jpeg()</code>, <code>tiff()</code> or <code>bmp()</code>
device but Cairo is unsupported (so <code>type = "Xlib"</code>
is tried instead).
</p>
</li>
<li> <p><code>grSoftVersion()</code> now reports the versions of FreeType
and FontConfig if they are used directly (not <em>via</em> Pango),
as is most commonly done on macOS.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li><p> The <em>standalone</em> ‘<span class="file">libRmath</span>’ math library and <span class="rlang"><b>R</b></span>'s C
API now provide <code>log1pexp()</code> again as documented, and gain
<code>log1mexp()</code>.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li> <p><code>configure</code> checks for a program <code>pkgconf</code>
if program <code>pkg-config</code> is not found. These are now only
looked for on the path (like almost all other programs) so if needed
specify a full path to the command in <code>PKG_CONFIG</code>, for example
in file ‘<span class="file">config.site</span>’.
</p>
</li>
<li><p> C99 function <code>iswblank</code> is required – it was last seen
missing ca 2003 so the workaround has been removed.
</p>
</li>
<li><p> There are new <code>configure</code> options
<span class="option">--with-internal-iswxxxxx</span>,
<span class="option">--with-internal-towlower</span> and
<span class="option">--with-internal-wcwidth</span> which allows the system functions
for wide-character classification, case-switching and width
(<code>wcwidth</code> and <code>wcswidth</code>) to be replaced by internal
ones. The first has long been used on macOS, AIX (and Windows)
but this enables it to be unselected there and selected for other
platforms (it is the new default on Solaris). The second is new
in this version of <span class="rlang"><b>R</b></span> and is selected by default on macOS and
Solaris. The third has long been the default and remains so as it
contains customizations for East Asian languages.
</p>
<p>System versions of these functions are often minimally implemented
(sometimes only for ASCII characters) and may not cover the full
range of Unicode points: for example Solaris (and Windows) only
cover the Basic Multilingual Plane.
</p>
</li>
<li><p> Cairo installations without X11 are more likely to be
detected by <code>configure</code>, when the file-based Cairo
graphics devices will be available but not <code>X11(type =
"cairo")</code>.
</p>
</li>
<li><p> There is a new <code>configure</code> option
<span class="option">--with-static-cairo</span> which is the default on macOS. This
should be used when only static cairo (and where relevant, Pango)
libraries are available.
</p>
</li>
<li> <p>Cairo-based graphics devices on platforms without Pango but
with FreeType/FontConfig will make use of the latter for font selection.
</p>
</li></ul>
<h4>LINK-TIME OPTIMIZATION on a UNIX-ALIKE</h4>
<ul>
<li><p> Configuring with flag <span class="option">--enable-lto=R</span> now also uses
<abbr>LTO</abbr> when installing the recommended packages.
</p>
</li>
<li> <p><code>R CMD INSTALL</code> and <code>R CMD SHLIB</code> have a new
flag <span class="option">--use-LTO</span> to use <abbr>LTO</abbr> when compiling code, for use
with <span class="rlang"><b>R</b></span> configured with <span class="option">--enable-lto=R</span>. For <span class="rlang"><b>R</b></span>
configured with <span class="option">--enable-lto</span>, they have the new flag
<span class="option">--no-use-LTO</span>.
</p>
<p>Packages can opt in or out of <abbr>LTO</abbr> compilation <em>via</em> a
‘<span class="samp">⁠UseLTO⁠</span>’ field in the ‘<span class="file">DESCRIPTION</span>’ file. (As usual this
can be overridden by the command-line flags.)
</p>
</li></ul>
<h4>BUILDING R on Windows</h4>
<ul>
<li><p> for GCC <code class="reqn">\ge</code> 8, <code>FC_LEN_T</code> is defined in
‘<span class="file">config.h</span>’ and hence character lengths are passed from C to
Fortran in <em>inter alia</em> BLAS and LAPACK calls.
</p>
</li>
<li><p> There is a new text file
‘<span class="file">src/gnuwin32/README.compilation</span>’, which outlines how C/Fortran
code compilation is organized and documents new features:
</p>
<ul>
<li> <p><span class="rlang"><b>R</b></span> can be built with Link-Time Optimization with a
suitable compiler – doing so with GCC 9.2 showed several
inconsistencies which have been corrected.
</p>
</li>
<li><p> There is support for cross-compiling the C and Fortran
code in <span class="rlang"><b>R</b></span> and standard packages on suitable (Linux) platforms.
This is mainly intended to allow developers to test later
versions of compilers – for example using GCC 9.2 or 10.x has
detected issues that GCC 8.3 in Rtools40 does not.
</p>
</li>
<li><p> There is experimental support for cross-building <span class="rlang"><b>R</b></span>
packages with C, C++ and/or Fortran code.
</p>
</li></ul>
</li>
<li><p> The R installer can now be optionally built to support a single
architecture (only 64-bit or only 32-bit).
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> The default C++ standard has been changed to C++14 where
available (which it is on all currently checked platforms): if not
(as before) C++11 is used if available otherwise C++ is not
supported.
</p>
<p>Packages which specify C++11 will still be installed using C++11.
</p>
<p>C++14 compilers may give deprecation warnings, most often for
<code>std::random_shuffle</code> (deprecated in C++14 and removed in
C++17). Either specify C++11 (see ‘Writing R Extensions’)
or modernize the code and if needed specify C++14. The latter has
been supported since <span class="rlang"><b>R</b></span> 3.4.0 so the package's ‘<span class="file">DESCRIPTION</span>’
would need to include something like
</p>
<pre>
Depends: R (>= 3.4)
</pre>
</li></ul>
<h4>PACKAGE INSTALLATION on Windows</h4>
<ul>
<li> <p><code>R CMD INSTALL</code> and <code>R CMD SHLIB</code> make use
of their flag <span class="option">--use-LTO</span> when the ‘<span class="samp">⁠LTO_OPT⁠</span>’ make
macro is set in file ‘<span class="file">etc/${R_ARCH}/Makeconf</span>’ or in a
personal/site ‘<span class="file">Makevars</span>’ file. (For details see
‘Writing R Extensions’ §4.5.)
</p>
<p>This provides a valuable check on code consistency. It does work
with GCC 8.3 as in Rtools40, but that does not detect everything
the <abbr>CRAN</abbr> checks with current GCC do.
</p>
</li></ul>
<h4>PACKAGE INSTALLATION on macOS</h4>
<ul>
<li><p> The default personal library directory on builds with
<span class="option">--enable-aqua</span> (including <abbr>CRAN</abbr> builds) now
differs by CPU type, one of
</p>
<pre>
~/Library/R/x86_64/x.y/library
~/Library/R/arm64/x.y/library
</pre>
<p>This uses the CPU type <span class="rlang"><b>R</b></span> (and hence the packages) were built for,
so when a ‘<span class="samp">⁠x86_64⁠</span>’ build of R is run under Rosetta emulation on
an ‘<span class="samp">⁠arm64⁠</span>’ Mac, the first is used.
</p>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> can now scan package functions for
bogus <code>return</code> statements, which were possibly intended as
<code>return()</code> calls (wish of <a href="https://bugs.R-project.org/show_bug.cgi?id=17180">PR#17180</a>, patch by Sebastian
Meyer). This check can be activated via the new environment
variable <span class="env">_R_CHECK_BOGUS_RETURN_</span>, true for <code>--as-cran</code>.
</p>
</li>
<li> <p><code>R CMD build</code> omits tarballs and binaries of
previous builds from the top-level package directory.
(<a href="https://bugs.R-project.org/show_bug.cgi?id=17828">PR#17828</a>, patch by Sebastian Meyer.)
</p>
</li>
<li> <p><code>R CMD check</code> now runs sanity checks on the use of
‘<span class="samp">⁠LazyData⁠</span>’, for example that a ‘<span class="file">data</span>’ directory is
present and that ‘<span class="samp">⁠LazyDataCompression⁠</span>’ is not specified
without ‘<span class="samp">⁠LazyData⁠</span>’ and has a documented value. For packages
with large LazyData databases without specifying
‘<span class="samp">⁠LazyDataCompression⁠</span>’, there is a reference to the code given
in ‘Writing R Extensions’ §1.1.6 to test the choice of
compression (as in all the <abbr>CRAN</abbr> packages tested a
non-default method was preferred).
</p>
</li>
<li> <p><code>R CMD build</code> removes ‘<span class="samp">⁠LazyData⁠</span>’ and
‘<span class="samp">⁠LazyDataCompression⁠</span>’ fields from the ‘<span class="file">DESCRIPTION</span>’ file
of packages without a ‘<span class="file">data</span>’ directory.
</p>
</li></ul>
<h4>ENCODING-RELATED CHANGES</h4>
<ul>
<li><p> The parser now treats ‘<span class="samp">⁠\Unnnnnnnn⁠</span>’ escapes larger than
the upper limit for Unicode points (‘<span class="samp">⁠\U10FFFF⁠</span>’) as an error
as they cannot be represented by valid UTF-8.
</p>
<p>Where such escapes are used for outputting non-printable
(including unassigned) characters, 6 hex digits are used (rather
than 8 with leading zeros). For clarity, braces are used, for
example ‘<span class="samp">⁠\U{0effff}⁠</span>’.
</p>
</li>
<li><p> The parser now looks for non-ASCII spaces on Solaris (as
previously on most other OSes).
</p>
</li>
<li><p> There are warnings (including from the parser) on the use of
unpaired surrogate Unicode points such as ‘<span class="samp">⁠\uD834⁠</span>’. (These
cannot be converted to valid UTF-8.)
</p>
</li>
<li><p> Functions <code>nchar()</code>, <code>tolower()</code>, <code>toupper()</code>
and <code>chartr()</code> and those using regular expressions have more
support for inputs with a marked Latin-1 encoding.
</p>
</li>
<li><p> The character-classification functions used (by default) to
replace the system <code>iswxxxxx</code> functions on Windows, macOS and
AIX have been updated to Unicode 13.0.0.
</p>
<p>The character-width tables have been updated to include new
assignments in Unicode 13.0.0. This included treating all control
characters as having zero width.
</p>
</li>
<li><p> The code for evaluating default (extended) regular
expressions now uses the same character-classification functions
as the rest of <span class="rlang"><b>R</b></span> (previously they differed on Windows, macOS and
AIX).
</p>
</li>
<li><p> There is a build-time option to replace the system's
wide-character <code>wctrans</code> C function by tables shipped with
<span class="rlang"><b>R</b></span>: use <code>configure</code> option
<span class="option">--with-internal-towlower</span> or (on Windows)
‘<span class="samp">⁠-DUSE_RI18N_CASE⁠</span>’ in ‘<span class="samp">⁠CFLAGS⁠</span>’ when building <span class="rlang"><b>R</b></span>. This
may be needed to allow <code>tolower()</code> and <code>toupper()</code> to
work with Unicode characters beyond the Basic Multilingual Plane
where not supported by system functions (e.g. on Solaris where it
is the new default).
</p>
</li>
<li><p> R is more careful when truncating UTF-8 and other multi-byte
strings that are too long to be printed, passed to the system or
libraries or placed into an internal buffer. Truncation will no
longer produce incomplete multibyte characters.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> Function <code>plclust()</code> from the package <span class="pkg">stats</span> and
<code>package.dependencies()</code>, <code>pkgDepends()</code>,
<code>getDepList()</code>, <code>installFoundDepends()</code>, and
<code>vignetteDepends()</code> from package <span class="pkg">tools</span> are defunct.
</p>
</li>
<li><p> Defunct functions <code>checkNEWS()</code> and <code>readNEWS()</code> from
package <span class="pkg">tools</span> and <code>CRAN.packages()</code> from <span class="pkg">utils</span> have
been removed.
</p>
</li>
<li> <p><code>R CMD config CXXCPP</code> is defunct (it was deprecated
in <span class="rlang"><b>R</b></span> 3.6.2).
</p>
</li>
<li> <p><code>parallel::detectCores()</code> drops support for IRIX
(retired in 2013).
</p>
</li>
<li><p> The <code>LINPACK</code> argument to <code>chol.default()</code>,
<code>chol2inv()</code>, <code>solve.default()</code> and <code>svd()</code> has
been defunct since <span class="rlang"><b>R</b></span> 3.1.0. It was silently ignored up to <span class="rlang"><b>R</b></span>
4.0.3 but now gives an error.
</p>
</li>
<li><p> Subsetting/indexing, such as <code>ddd[*]</code> or <code>ddd$x</code>
on a <code>DOTSXP</code> (dot-dot-dot) object <code>ddd</code> has been
disabled; it worked by accident only and was undocumented.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Many more C-level allocations (mainly by <code>malloc</code> and
<code>strdup</code>) are checked for success with suitable alternative
actions.
</p>
</li>
<li><p> Bug fix for <code>replayPlot()</code>; this was turning off
graphics engine display list recording if a recorded plot was
replayed in the same session. The impact of the bug became
visible if resize the device after replay OR if attempted another
<code>savePlot()</code> after replay (empty display list means empty
screen on resize or empty saved plot).
</p>
</li>
<li> <p><code>R CMD check</code> etc now warn when a package exports
non-existing S4 classes or methods, also in case of no
“methods” presence. (Reported by Alex Bertram;
reproducible example and patch by Sebastian Meyer in <a href="https://bugs.R-project.org/show_bug.cgi?id=16662">PR#16662</a>.)
</p>
</li>
<li> <p><code>boxplot()</code> now also accepts <code>call</code>s for labels such
as <code>ylab</code>, the same as <code>plot()</code>.
(Reported by Marius Hofert.)
</p>
</li>
<li><p> The help page for <code>xtabs()</code> now correctly states that
<code>addNA</code> is setting <code>na.action = na.pass</code> among others.
(Reported as <a href="https://bugs.R-project.org/show_bug.cgi?id=17770">PR#17770</a> by Thomas Soeiro.)
</p>
</li>
<li><p> The <code>R CMD check <pkg></code> gives a longer and more
comprehensible message when ‘<span class="file">DESCRIPTION</span>’ misses dependencies,
e.g., in <code>Imports:</code>. (Thanks to the contributors of <a href="https://bugs.R-project.org/show_bug.cgi?id=17179">PR#17179</a>.)
</p>
</li>
<li> <p><code>update.default()</code> now calls the generic <code>update()</code>
on the formula to work correctly for models with extended formulas.
(As reported and suggested by Neal Fultz in <a href="https://bugs.R-project.org/show_bug.cgi?id=17865">PR#17865</a>.)
</p>
</li>
<li><p> The horizontal position of leaves in a dendrogram is now
correct also with <code>center = FALSE</code>. (<a href="https://bugs.R-project.org/show_bug.cgi?id=14938">PR#14938</a>, patch from
Sebastian Meyer.)
</p>
</li>
<li> <p><code>all.equal.POSIXt()</code> no longer warns about and
subsequently ignores inconsistent <code>"tzone"</code> attributes, but
describes the difference in its return value (<a href="https://bugs.R-project.org/show_bug.cgi?id=17277">PR#17277</a>).
This check can be disabled <em>via</em> the new argument
<code>check.tzone = FALSE</code> as suggested by Sebastian Meyer.
</p>
</li>
<li> <p><code>as.POSIXct()</code> now populates the <code>"tzone"</code>
attribute from its <code>tz</code> argument when <code>x</code> is a logical
vector consisting entirely of <code>NA</code> values.
</p>
</li>
<li> <p><code>x[[2^31]] <- v</code> now works. (Thanks to the report and
patch by Suharto Anggono in <a href="https://bugs.R-project.org/show_bug.cgi?id=17330">PR#17330</a>.)
</p>
</li>
<li><p> In log-scale graphics, <code>axis()</code> ticks and label positions
are now computed more carefully and symmetrically in their range,
typically providing <em>more</em> ticks, fulfilling wishes
in <a href="https://bugs.R-project.org/show_bug.cgi?id=17936">PR#17936</a>. The change really corresponds to an improved
<code>axisTicks()</code> (package <span class="pkg">grDevices</span>), potentially influencing
<span class="pkg">grid</span> and <span class="pkg">lattice</span>, for example.
</p>
</li>
<li> <p><code>qnorm(<very large negative>, log.p=TRUE)</code> is now correct
to at least five digits where it was catastrophically wrong,
previously.
</p>
</li>
<li> <p><code>sum(df)</code> and similar <code>"Summary"</code>- and
<code>"Math"</code>-group member functions now work for data frames
<code>df</code> with <code>logical</code> columns, notably also of zero
rows. (Reported to R-devel by Martin “b706”.)
</p>
</li>
<li> <p><code>unsplit()</code> had trouble with tibbles due to unsound use of
<code>rep(NA, len)</code>-indexing, which should use <code>NA_integer_</code>
(Reported to R-devel by Mario Annau.)
</p>
</li>
<li> <p><code>pnorm(x, log.p = TRUE)</code> underflows to <code>-Inf</code> slightly
later.
</p>
</li>
<li> <p><code>show(<hidden S4 generic>)</code> prints better and without
quotes for non-hidden S4 generics.
</p>
</li>
<li> <p><code>read.table()</code> and relatives treated an "NA" column name as
missing when <code>check.names = FALSE</code> <a href="https://bugs.R-project.org/show_bug.cgi?id=18007">PR#18007</a>.
</p>
</li>
<li><p> Parsing strings containing UTF-16 surrogate pairs such as
<code>"\uD834\uDD1E"</code> works better on some (uncommon) platforms.
<code>sprintf("%X", utf8ToInt("\uD834\uDD1E"))</code> should now give
<code>"1D11E"</code> on all platforms.
</p>
</li>
<li> <p><code>identical(x,y)</code> is no longer true for differing
<code>DOTSXP</code> objects, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18032">PR#18032</a>.
</p>
</li>
<li> <p><code>str()</code> now works correctly for <code>DOTSXP</code> and related
exotics, even when these are doomed.
</p>
<p>Additionally, it no longer fails for <code>list</code>s with a <code>class</code> and
“irregular” method definitions such that e.g. <code>lapply(*)</code> will
necessarily fail, as currently for different <a href="https://CRAN.R-project.org/package=igraph"><span class="pkg">igraph</span></a> objects.
</p>
</li>
<li><p> Message translation domains, e.g., for errors and warnings, are
now correctly determined also when e.g., a <span class="pkg">base</span> function is
called from “top-level” function (i.e., defined in <code>globalenv()</code>),
thanks to a patch from Joris Goosen fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17998">PR#17998</a>.
</p>
</li>
<li><p> Too long lines in environment files (e.g., ‘<span class="file">Renviron</span>’) no
longer crash <span class="rlang"><b>R</b></span>. This limit has been increased to 100,000 bytes.
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18001">PR#18001</a>.)
</p>
</li>
<li><p> There is a further workaround for FreeType giving
incorrect italic font faces with cairo-based graphics devices on
macOS.
</p>
</li>
<li> <p><code>add_datalist(*, force = TRUE)</code> (from package
<span class="pkg">tools</span>) now actually updates an existing ‘<span class="file">data/datalist</span>’
file for new content. (Thanks to a report and patch by Sebastian
Meyer in <a href="https://bugs.R-project.org/show_bug.cgi?id=18048">PR#18048</a>.)
</p>
</li>
<li> <p><code>cut.Date()</code> and <code>cut.POSIXt()</code> could produce an
empty last interval for <code>breaks = "months"</code> or <code>breaks =
"years"</code>. (Reported as <a href="https://bugs.R-project.org/show_bug.cgi?id=18053">PR#18053</a> by Christopher Carbone.)
</p>
</li>
<li><p> Detection of the encoding of ‘regular’ macOS locales
such as ‘<span class="samp">⁠en_US⁠</span>’ (which is UTF-8) had been broken by a macOS
change: fortunately these are now rarely used with
‘<span class="samp">⁠en_US.UTF-8⁠</span>’ being preferred.
</p>
</li>
<li> <p><code>sub()</code> and <code>gsub(pattern, repl, x, *)</code> now keep
attributes of <code>x</code> such as <code>names()</code> also when
<code>pattern</code> is <code>NA</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18079">PR#18079</a>).
</p>
</li>
<li><p> Time differences (<code>"difftime"</code> objects) get a replacement
and a <code>rep()</code> method to keep <code>"units"</code> consistent.
(Thanks to a report and patch by Nicolas Bennett in <a href="https://bugs.R-project.org/show_bug.cgi?id=18066">PR#18066</a>.)
</p>
</li>
<li><p> The <code style="white-space: pre;">⁠\RdOpts⁠</code> macro, setting defaults for <code style="white-space: pre;">⁠\Sexpr⁠</code>
options in an Rd file, had been ineffective since <span class="rlang"><b>R</b></span> 2.12.0: it
now works again.
(Thanks to a report and patch by Sebastian Meyer in <a href="https://bugs.R-project.org/show_bug.cgi?id=18073">PR#18073</a>.)
</p>
</li>
<li> <p><code>mclapply</code> and <code>pvec</code> no longer accidentally terminate
parallel processes started before by <code>mcparallel</code> or related
calls in package <span class="pkg">parallel</span> (<a href="https://bugs.R-project.org/show_bug.cgi?id=18078">PR#18078</a>).
</p>
</li>
<li> <p><code>grep</code> and other functions for evaluating (extended)
regular expressions handle in Unicode also strings not explicitly
flagged UTF-8, but flagged native when running in UTF-8 locale.
</p>
</li>
<li><p> Fixed a crash in <code>fifo</code> implementation on Windows
(<a href="https://bugs.R-project.org/show_bug.cgi?id=18031">PR#18031</a>).
</p>
</li>
<li><p> Binary mode in <code>fifo</code> on Windows is now properly detected
from argument <code>open</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=15600">PR#15600</a>, <a href="https://bugs.R-project.org/show_bug.cgi?id=18031">PR#18031</a>).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.5</h3>
<h4>BUG FIXES</h4>
<ul>
<li><p> The change to the internal table in <span class="rlang"><b>R</b></span> 4.0.4 for
<code>iswprint</code> has been reverted: it contained some errors in
printability of ‘East Asian’ characters.
</p>
</li>
<li><p> For packages using ‘<span class="samp">⁠LazyData⁠</span>’, <code>R CMD build</code>
ignored the <span class="option">--resave-data</span> option and the
‘<span class="samp">⁠BuildResaveData⁠</span>’ field of the ‘<span class="file">DESCRIPTION</span>’ file (in <span class="rlang"><b>R</b></span>
versions 4.0.0 to 4.0.4).
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.4</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> File ‘<span class="file">share/texmf/tex/latex/jss.cls</span>’ has been updated
to work with LaTeX versions since Oct 2020.
</p>
</li>
<li><p> Unicode character width tables (as used by
<code>nchar(, type = "w")</code>) have been updated to Unicode 12.1
by Brodie Gaslam (<a href="https://bugs.R-project.org/show_bug.cgi?id=17781">PR#17781</a>), including many emoji.
</p>
</li>
<li><p> The internal table for <code>iswprint</code> (used on Windows,
macOS and AIX) has been updated to include many recent Unicode
characters.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> If an external BLAS is specified by <span class="option">--with-blas=foo</span>
or <em>via</em> environment variable <span class="env">BLAS_LIBS</span> is not found,
this is now a configuration error. The previous behaviour was not
clear from the documentation: it was to continue the search as if
<span class="option">--with-blas=yes</span> was specified.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>all.equal(x,y)</code> now “sees” the two different
<code>NA</code>s in factors, thanks to Bill Dunlap and others in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17897">PR#17897</a>.
</p>
</li>
<li> <p><code>(~ NULL)[1]</code> and similar formula subsetting now works,
thanks to a report and patch by Henrik Bengtsson in <a href="https://bugs.R-project.org/show_bug.cgi?id=17935">PR#17935</a>.
Additionally, subsetting leaving an empty formula now works too,
thanks to suggestions by Suharto Anggono.
</p>
</li>
<li> <p><code>.traceback(n)</code> keeps source references again, as before
<span class="rlang"><b>R</b></span> 4.0.0, fixing a regression; introduced by the <a href="https://bugs.R-project.org/show_bug.cgi?id=17580">PR#17580</a>, reported
including two patch proposals by Brodie Gaslam.
</p>
</li>
<li> <p><code>unlist(plst, recursive=FALSE)</code> no longer drops content
for pairlists with list components, thanks to the report and patch
by Suharto Anggono in <a href="https://bugs.R-project.org/show_bug.cgi?id=17950">PR#17950</a>.
</p>
</li>
<li> <p><code>iconvlist()</code> now also works on MUSL based (Linux)
systems, from a report and patch suggestion by Wesley Chan in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17970">PR#17970</a>.
</p>
</li>
<li> <p><code>round()</code> and <code>signif()</code> no longer tolerate wrong
argument names, notably in 1-argument calls; reported by Shane
Mueller on R-devel (mailing list); later reported as <a href="https://bugs.R-project.org/show_bug.cgi?id=17976">PR#17976</a>.
</p>
</li>
<li> <p><code>.Machine</code> has <code>longdouble.*</code> elements only if
<code>capabilities("long.double")</code> is true, as documented.
(Previously they were included if the platform had <code>long
double</code> identical to <code>double</code>, as ARM does.)
</p>
</li>
<li> <p><code>p.adjust(numeric(), n=0)</code> now works, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18002">PR#18002</a>.
</p>
</li>
<li> <p><code>identical(x,y)</code> no longer prints "Unknown Type .." for
<code>typeof(x) == "..."</code> objects.
</p>
</li>
<li><p> Fix (auto-)<code>print()</code>ing of named complex vectors, see
<a href="https://bugs.R-project.org/show_bug.cgi?id=17868">PR#17868</a> and <a href="https://bugs.R-project.org/show_bug.cgi?id=18019">PR#18019</a>.
</p>
</li>
<li> <p><code>all.equal(<language>, <...>)</code> now works, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18029">PR#18029</a>.
</p>
</li>
<li> <p><code>as.data.frame.list(L, row.names=NULL)</code> now behaves in line
with <code>data.frame()</code>, disregarding names of components of
<code>L</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=18034">PR#18034</a>, reported by Kevin Tappe.
</p>
</li>
<li> <p><code>checkRdaFiles(ff)$version</code> is now correct also when
<code>ff</code> contains files of different versions, thanks to a report
and patch from Sebastian Meyer in <a href="https://bugs.R-project.org/show_bug.cgi?id=18041">PR#18041</a>.
</p>
</li>
<li><p> macOS: Quartz device live drawing could fail (no plot is shown)
if the system changes the drawing context after view update (often
the case since macOS Big Sur). System log may show
"CGContextDelegateCreateForContext: invalid context" error.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.3</h3>
<h4>NEW FEATURES</h4>
<ul>
<li><p> On platforms using <code>configure</code> option
<span class="option">--with-internal-tzcode</span>, additional values
<code>"internal"</code> and (on macOS only) <code>"macOS"</code> are accepted
for the environment variable <span class="env">TZDIR</span>. (See <code>?TZDIR</code>.)
</p>
<p>On macOS, <code>"macOS"</code> is used by default if the system timezone
database is a newer version than that in the <span class="rlang"><b>R</b></span> installation.
</p>
</li>
<li><p> When <code>install.packages(type = "source")</code> fails to find
a package in a repository it mentions package versions which are
excluded by their <span class="rlang"><b>R</b></span> version requirement and links to hints on
why a package might not be found.
</p>
</li>
<li><p> The default value for <code>options("timeout")</code> can be set
from environment variable <span class="env">R_DEFAULT_INTERNET_TIMEOUT</span>, still
defaulting to 60 (seconds) if that is not set or invalid.
</p>
<p>This may be needed when child <span class="rlang"><b>R</b></span> processes are doing downloads,
for example during the installation of source packages which
download jars or other forms of data.
</p>
</li></ul>
<h4>LINK-TIME OPTIMIZATION on a UNIX-ALIKE</h4>
<ul>
<li><p> There is now support for parallelized Link-Time Optimization
(<abbr>LTO</abbr>) with GCC and for ‘thin’ <abbr>LTO</abbr> with
<code>clang</code> <em>via</em> setting the ‘<span class="samp">⁠LTO⁠</span>’ macro.
</p>
</li>
<li><p> There is support for setting a different <abbr>LTO</abbr> flag for the
Fortran compiler, including to empty when mixing <code>clang</code>
and <code>gfortran</code> (as on macOS). See file ‘<span class="file">config.site</span>’.
</p>
</li>
<li><p> There is a new ‘<span class="samp">⁠LTO_LD⁠</span>’ macro to set linker options for
<abbr>LTO</abbr> compilation, for example to select an alternative linker or
to parallelize thin <abbr>LTO</abbr>.
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> The <code>LINPACK</code> argument to <code>chol.default()</code>,
<code>chol2inv()</code>, <code>solve.default()</code> and <code>svd()</code> has been
defunct since <span class="rlang"><b>R</b></span> 3.1.0. Using it now gives a warning which will
become an error in <span class="rlang"><b>R</b></span> 4.1.0.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> The code mitigating stack overflow with PCRE regexps on very
long strings is enabled for PCRE2 < 10.30 also when <abbr>JIT</abbr> is enabled,
since stack overflows have been seen in that case.
</p>
</li>
<li><p> Fix to correctly show the group labels in <code>dotchart()</code>
(which where lost in the <code>ylab</code> improvement for <span class="rlang"><b>R</b></span> 4.0.0).
</p>
</li>
<li> <p><code>addmargins(*, ..)</code> now also works when <code>fn()</code> is a
local function, thanks to bug report and patch <a href="https://bugs.R-project.org/show_bug.cgi?id=17124">PR#17124</a> from Alex
Bertram.
</p>
</li>
<li> <p><code>rank(x)</code> and hence <code>sort(x)</code> now work when <code>x</code>
is an object (as per <code>is.object(x)</code>) of type <code>"raw"</code>
<em>and</em> provides a valid <code>`[`</code> method, e.g., for
<code>gmp::as.bigz(.)</code> numbers.
</p>
</li>
<li> <p><code>chisq.test(*, simulate.p.value=TRUE)</code> and
<code>r2dtable()</code> now work correctly for large table entries (in the
millions). Reported by Sebastian Meyer and investigated by more
helpers in <a href="https://bugs.R-project.org/show_bug.cgi?id=16184">PR#16184</a>.
</p>
</li>
<li><p> Low-level socket read/write operations have been fixed to
correctly signal communication errors. Previously, such errors could
lead to a segfault due to invalid memory access. Reported
and debugged by Dmitriy Selivanov in <a href="https://bugs.R-project.org/show_bug.cgi?id=17850">PR#17850</a>.
</p>
</li>
<li> <p><code>quantile(x, pr)</code> works more consistently for <code>pr</code>
values slightly outside [0,1], thanks to Suharto Anggono's <a href="https://bugs.R-project.org/show_bug.cgi?id=17891">PR#17891</a>.
</p>
<p>Further, <code>quantile(x, prN, names=FALSE)</code> now works even when
<code>prN</code> contains <code>NA</code>s, thanks to Anggono's <a href="https://bugs.R-project.org/show_bug.cgi?id=17892">PR#17892</a>.
Ditto for ordered factors or <code>Date</code> objects
when <code>type = 1</code> or <code>3</code>, thanks to <a href="https://bugs.R-project.org/show_bug.cgi?id=17899">PR#17899</a>.
</p>
</li>
<li><p> Internet access based on libcurl, including
<code>curlGetHeaders()</code>, was not respecting the <code>"timeout"</code>
option. If this causes unanticipated timeouts, consider
increasing the default by setting <span class="env">R_DEFAULT_INTERNET_TIMEOUT</span>.
</p>
</li>
<li> <p><code>as.Date(<char>)</code> now also works with an initial
<code>""</code>, thanks to Michael Chirico's <a href="https://bugs.R-project.org/show_bug.cgi?id=17909">PR#17909</a>.
</p>
</li>
<li> <p><code>isS3stdGeneric(f)</code> now detects an S3 generic also when
it is <code>trace()</code>d, thanks to Gabe Becker's <a href="https://bugs.R-project.org/show_bug.cgi?id=17917">PR#17917</a>.
</p>
</li>
<li> <p><code>R_allocLD()</code> has been fixed to return memory aligned for
long double type <a href="https://bugs.R-project.org/show_bug.cgi?id=16534">PR#16534</a>.
</p>
</li>
<li> <p><code>fisher.test()</code> no longer segfaults when called again after
its internal stack has been exceeded <a href="https://bugs.R-project.org/show_bug.cgi?id=17904">PR#17904</a>.
</p>
</li>
<li><p> Accessing a long vector represented by a compact integer
sequence no longer segfaults (reported and debugged by Hugh
Parsonage).
</p>
</li>
<li> <p><code>duplicated()</code> now works also for strings with multiple
encodings inside a single vector <a href="https://bugs.R-project.org/show_bug.cgi?id=17809">PR#17809</a>.
</p>
</li>
<li> <p><code>phyper(11, 15, 0, 12, log.p=TRUE)</code> no longer gives
<code>NaN</code>; reported as <a href="https://bugs.R-project.org/show_bug.cgi?id=17271">PR#17271</a> by Alexey Stukalov.
</p>
</li>
<li><p> Fix incorrect calculation in <code>logLik.nls()</code> <a href="https://bugs.R-project.org/show_bug.cgi?id=16100">PR#16100</a>, patch
from Sebastian Meyer.
</p>
</li>
<li><p> A very old bug could cause a segfault in <code>model.matrix()</code>
when terms involved logical variables. Part of <a href="https://bugs.R-project.org/show_bug.cgi?id=17879">PR#17879</a>.
</p>
</li>
<li> <p><code>model.frame.default()</code> allowed <code>data = 1</code>, leading to
involuntary variable capture (rest of <a href="https://bugs.R-project.org/show_bug.cgi?id=17879">PR#17879</a>).
</p>
</li>
<li> <p><code>tar()</code> no longer skips non-directory files, thanks to a
patch by Sebastian Meyer, fixing the remaining part of <a href="https://bugs.R-project.org/show_bug.cgi?id=16716">PR#16716</a>.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.2</h3>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R CMD check</code> skips vignette re-building (with a
warning) if the ‘<span class="samp">⁠VignetteBuilder⁠</span>’ package(s) are not available.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li><p> Paths with non-ASCII characters caused problems for
package loading on Windows <a href="https://bugs.R-project.org/show_bug.cgi?id=17833">PR#17833</a>.
</p>
</li>
<li><p> Using <span class="pkg">tcltk</span> widgets no longer crashes R on
Windows.
</p>
</li>
<li> <p><code>source(*, echo=TRUE)</code> no longer fails in some cases with
empty lines; reported by Bill Dunlap in <a href="https://bugs.R-project.org/show_bug.cgi?id=17769">PR#17769</a>.
</p>
</li>
<li> <p><code>on.exit()</code> now correctly matches named arguments, thanks
to <a href="https://bugs.R-project.org/show_bug.cgi?id=17815">PR#17815</a> (including patch) by Brodie Gaslam.
</p>
</li>
<li> <p><code>regexpr(*, perl=TRUE)</code> no longer returns incorrect
positions into text containing characters outside of the Unicode
Basic Multilingual Plane on Windows.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.1</h3>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>paste()</code> and <code>paste0()</code> gain a new optional
argument <code>recycle0</code>. When set to true, zero-length
arguments are recycled leading to <code>character(0)</code> after the
<code>sep</code>-concatenation, i.e., to the empty string <code>""</code> if
<code>collapse</code> is a string and to the zero-length value
<code>character(0)</code> when <code>collapse = NULL</code>.
</p>
<p>A package whose code uses this should depend on ‘<span class="samp">⁠R (>= 4.0.1)⁠</span>’.
</p>
</li>
<li><p> The <code>summary(<warnings>)</code> method now maps the counts
correctly to the warning messages.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>aov(frml, ...)</code> now also works where the <code>formula</code>
deparses to more than 500 characters, thanks to a report and patch
proposal by Jan Hauffa.
</p>
</li>
<li><p> Fix a dozen places (code, examples) as <code>Sys.setlocale()</code>
returns the new rather than the previous setting.
</p>
</li>
<li><p> Fix for adding two complex <span class="pkg">grid</span> units via <code>sum()</code>.
Thanks to Gu Zuguang for the report and Thomas Lin Pedersen for
the patch.
</p>
</li>
<li><p> Fix <code>parallel::mclapply(..., mc.preschedule=FALSE)</code>
to handle raw vector results correctly. <a href="https://bugs.R-project.org/show_bug.cgi?id=17779">PR#17779</a>
</p>
</li>
<li><p> Computing the <code>base</code> value, i.e., 2, “everywhere”,
now uses <code>FLT_RADIX</code>, as the original ‘<span class="file">machar</span>’ code looped
indefinitely on the ppc64 architecture for the <code>longdouble</code> case.
</p>
</li>
<li><p> In <span class="rlang"><b>R</b></span> 4.0.0, <code>sort.list(x)</code> when <code>is.object(x)</code> was
true, e.g., for <code>x <- I(letters)</code>, was accidentally using
<code>method = "radix"</code>. Consequently, e.g., <code>merge(<data.frame>)</code>
was much slower than previously; reported in <a href="https://bugs.R-project.org/show_bug.cgi?id=17794">PR#17794</a>.
</p>
</li>
<li> <p><code>plot(y ~ x, ylab = quote(y[i]))</code> now works, as e.g., for
<code>xlab</code>; related to <a href="https://bugs.R-project.org/show_bug.cgi?id=10525">PR#10525</a>.
</p>
</li>
<li> <p><code>parallel::detect.cores(all.tests = TRUE)</code> tries a
matching OS name before the other tests (which were intended only
for unknown OSes).
</p>
</li>
<li><p> Parse data for raw strings is now recorded correctly. Reported
by Gabor Csardi.
</p>
</li></ul>
<h3><img src="../help/figures/../../html/Rlogo.svg" class="toplogo" alt="[R logo]"> CHANGES IN R 4.0.0</h3>
<h4>SIGNIFICANT USER-VISIBLE CHANGES</h4>
<ul>
<li><p> Packages need to be (re-)installed under this version
(4.0.0) of <span class="rlang"><b>R</b></span>.
</p>
</li>
<li> <p><code>matrix</code> objects now also inherit from class
<code>"array"</code>, so e.g., <code>class(diag(1))</code> is <code>c("matrix",
"array")</code>. This invalidates code incorrectly assuming that
<code>class(matrix_obj))</code> has length one.
</p>
<p>S3 methods for class <code>"array"</code> are now dispatched for
<code>matrix</code> objects.
</p>
</li>
<li><p> There is a new syntax for specifying <em>raw</em> character
constants similar to the one used in C++: <code>r"(...)"</code> with
<code>...</code> any character sequence not containing the sequence
‘<span class="samp">⁠)"⁠</span>’.
This makes it easier to write strings that contain backslashes or
both single and double quotes. For more details see <code>?Quotes</code>.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> now uses a ‘<span class="samp">⁠stringsAsFactors = FALSE⁠</span>’ default, and
hence by default no longer converts strings to factors in calls
to <code>data.frame()</code> and <code>read.table()</code>.
</p>
<p>A large number of packages relied on the previous behaviour and
so have needed/will need updating.
</p>
</li>
<li><p> The <code>plot()</code> S3 generic function is now in package
<span class="pkg">base</span> rather than package <span class="pkg">graphics</span>, as it is
reasonable to have methods that do not use the <span class="pkg">graphics</span>
package. The generic is currently re-exported from the
<span class="pkg">graphics</span> namespace to allow packages importing it from
there to continue working, but this may change in future.
</p>
<p>Packages which define S4 generics for <code>plot()</code> should be
re-installed and package code using such generics from other
packages needs to ensure that they are imported rather than rely
on their being looked for on the search path (as in a namespace,
the base namespace has precedence over the search path).
</p>
</li></ul>
<h4>REFERENCE COUNTING</h4>
<ul>
<li><p> Reference counting is now used instead of the <code>NAMED</code>
mechanism for determining when objects can be safely mutated in
base C code. This reduces the need for copying in some cases and
should allow further optimizations in the future. It should
help make the internal code easier to maintain.
</p>
<p>This change is expected to have almost no impact on packages
using supported coding practices in their C/C++ code.
</p>
</li></ul>
<h4>MIGRATION TO PCRE2</h4>
<ul>
<li><p> This version of <span class="rlang"><b>R</b></span> is built against the PCRE2 library for
Perl-like regular expressions, if available. (On non-Windows
platforms PCRE1 can optionally be used if PCRE2 is not available
at build time.) The version of PCRE in use can be obtained
<em>via</em> <code>extSoftVersion()</code>: PCRE1 (formerly known as
‘PCRE’) has versions <= 8, PCRE2 versions >= 10.
</p>
</li>
<li><p> Making PCRE2 available when building <span class="rlang"><b>R</b></span> from source is
strongly recommended (preferably version 10.30 or later) as PCRE1
is no longer developed: version 8.44 is
‘likely to be the final release’.
</p>
</li>
<li><p> PCRE2 reports errors for some regular expressions that were
accepted by PCRE1. A hyphen now has to be escaped in a character
class to be interpreted as a literal (unless first or last in the
class definition). ‘<span class="samp">⁠\R⁠</span>’, ‘<span class="samp">⁠\B⁠</span>’ and ‘<span class="samp">⁠\X⁠</span>’ are no
longer allowed in character classes (PCRE1 treated these as
literals).
</p>
</li>
<li><p> Option <code>PCRE_study</code> is no longer used with PCRE2, and is
reported as <code>FALSE</code> when that is in use.
</p>
</li></ul>
<h4>NEW FEATURES</h4>
<ul>
<li> <p><code>assertError()</code> and <code>assertWarning()</code> (in package
<span class="pkg">tools</span>) can now check for <em>specific</em> error or warning
classes <em>via</em> the new optional second argument <code>classes</code>
(which is not back compatible with previous use of an unnamed second
argument).
</p>
</li>
<li> <p><code>DF2formula()</code>, the utility for the data frame method of
<code>formula()</code>, now works without parsing and explicit evaluation,
starting from Suharto Anggono's suggestion in <a href="https://bugs.R-project.org/show_bug.cgi?id=17555">PR#17555</a>.
</p>
</li>
<li> <p><code>approxfun()</code> and <code>approx()</code> gain a new argument
<code>na.rm</code> defaulting to true. If set to false, missing
<code>y</code> values now propagate into the interpolated values.
</p>
</li>
<li><p> Long vectors are now supported as the <code>seq</code> argument of
a <code>for()</code> loop.
</p>
</li>
<li> <p><code>str(x)</code> gets a new <code>deparse.lines</code> option with a
default to speed it up when <code>x</code> is a large <code>call</code> object.
</p>
</li>
<li><p> The internal traceback object produced when an error is
signalled (<code>.Traceback</code>), now contains the <code>call</code>s
rather than the <em><code>deparse()</code>d</em> calls, deferring the
deparsing to the user-level functions <code>.traceback()</code> and
<code>traceback()</code>. This fulfils the wish of <a href="https://bugs.R-project.org/show_bug.cgi?id=17580">PR#17580</a>, reported
including two patch proposals by Brodie Gaslam.
</p>
</li>
<li> <p><code>data.matrix()</code> now converts character columns to
factors and from this to integers.
</p>
</li>
<li> <p><code>package.skeleton()</code> now explicitly lists all exports
in the ‘<span class="file">NAMESPACE</span>’ file.
</p>
</li>
<li><p> New function <code>.S3method()</code> to register S3 methods in R
scripts.
</p>
</li>
<li> <p><code>file.path()</code> has some support for file paths not in
the session encoding, e.g. with UTF-8 inputs in a
non-UTF-8 locale the output is marked as UTF-8.
</p>
</li>
<li><p> Most functions with file-path inputs will give an explicit
error if a file-path input in a marked encoding cannot be
translated (to the native encoding or in some cases on Windows to
UTF-8), rather than translate to a different file path using
escapes. Some (such as <code>dir.exists()</code>, <code>file.exists()</code>,
<code>file.access()</code>, <code>file.info()</code>, <code>list.files()</code>,
<code>normalizePath()</code> and <code>path.expand()</code>) treat this like
any other non-existent file, often with a warning.
</p>
</li>
<li><p> There is a new help document accessed by
<code>help("file path encoding")</code> detailing how file paths with
marked encodings are handled.
</p>
</li>
<li><p> New function <code>list2DF()</code> for creating data frames from
lists of variables.
</p>
</li>
<li> <p><code>iconv()</code> has a new option <code>sub = "Unicode"</code> to
translate UTF-8 input invalid in the ‘<span class="samp">⁠to⁠</span>’ encoding using
‘<span class="samp">⁠<U+xxxx>⁠</span>’ escapes.
</p>
</li>
<li><p> There is a new function <code>infoRDS()</code> providing information
about the serialization format of a serialized object.
</p>
</li>
<li><p> S3 method lookup now by default skips the elements of the
search path between the global and base environments.
</p>
</li>
<li><p> Added an argument <code>add_datalist(*, small.size = 0)</code> to
allow the creation of a ‘<span class="file">data/datalist</span>’ file even when the
total size of the data sets is small.
</p>
</li>
<li><p> The backquote function <code>bquote()</code> has a new argument
<code>splice</code> to enable splicing a computed list of values into an
expression, like <code>,@</code> in <abbr>LISP</abbr>'s backquote.
</p>
</li>
<li><p> The formula interface to <code>t.test()</code> and
<code>wilcox.test()</code> has been extended to handle one-sample and
paired tests.
</p>
</li>
<li><p> The <code>palette()</code> function has a new default set of
colours (which are less saturated and have better accessibility
properties). There are also some new built-in palettes, which
are listed by the new <code>palette.pals()</code> function. These
include the old default palette under the name <code>"R3"</code>. Finally,
the new <code>palette.colors()</code> function allows a subset of
colours to be selected from any of the built-in palettes.
</p>
</li>
<li> <p><code>n2mfrow()</code> gains an option <code>asp = 1</code> to specify the
aspect ratio, fulfilling the wish and extending the proposal of
Michael Chirico in <a href="https://bugs.R-project.org/show_bug.cgi?id=17648">PR#17648</a>.
</p>
</li>
<li><p> For <code>head(x, n)</code> and <code>tail()</code> the default and
other S3 methods notably for <em>vector</em> <code>n</code>, e.g. to get a
“corner” of a matrix, has been extended to <code>array</code>'s
of higher dimension thanks to the patch proposal by Gabe Becker in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17652">PR#17652</a>. Consequently, optional argument <code>addrownums</code> is
deprecated and replaced by the (more general) argument
<code>keepnums</code>. An invalid second argument <code>n</code> now leads
to typically more easily readable error messages.
</p>
</li>
<li><p> New function <code>.class2()</code> provides the full character
vector of class names used for S3 method dispatch.
</p>
</li>
<li><p> Printing <code>methods(..)</code> now uses a new <code>format()</code> method.
</p>
</li>
<li> <p><code>sort.list(x)</code> now works for non-atomic objects
<code>x</code> and <code>method = "auto"</code> (the default) or
<code>"radix"</code> in cases <code>order(x)</code> works, typically via a
<code>xtfrm()</code> method.
</p>
</li>
<li><p> Where they are available, <code>writeBin()</code> allows
long vectors.
</p>
</li>
<li><p> New function <code>deparse1()</code> produces one string, wrapping
<code>deparse()</code>, to be used typically in
<code>deparse1(substitute(*))</code>, e.g., to fix <a href="https://bugs.R-project.org/show_bug.cgi?id=17671">PR#17671</a>.
</p>
</li>
<li> <p><code>wilcox.test()</code> enhancements: In the (non-paired)
two-sample case, <code>Inf</code> values are treated as very large for
robustness consistency. If exact computations are used, the
result now has <code>"exact"</code> in the <code>method</code> element of its
return value. New arguments <code>tol.root</code> and
<code>digits.rank</code> where the latter may be used for stability to
treat very close numbers as ties.
</p>
</li>
<li> <p><code>readBin()</code> and <code>writeBin()</code> now report an error
for an invalid <code>endian</code> value. The affected code needs to be
fixed with care as the old undocumented behavior was to swap
endianness in such cases.
</p>
</li>
<li> <p><code>sequence()</code> is now an S3 generic with an internally
implemented default method, and gains arguments to generate more
complex sequences. Based on code from the <span class="pkg">S4Vectors</span>
Bioconductor package and the advice of Hervé Pagès.
</p>
</li>
<li> <p><code>print()</code>'s default method and many other methods (by
calling the default eventually and passing <code>...</code>) now make
use of a new optional <code>width</code> argument, avoiding the need for
the user to set and reset <code>options("width")</code>.
</p>
</li>
<li> <p><code>memDecompress()</code> supports the RFC 1952 format
(e.g. in-memory copies of <code>gzip</code>-compressed files)
as well as RFC 1950.
</p>
</li>
<li> <p><code>memCompress()</code> and <code>memDecompress()</code> support long
raw vectors for types <code>"gzip"</code> and <code>"zx"</code>.
</p>
</li>
<li> <p><code>sweep()</code> and <code>slice.index()</code> can now use names
of dimnames for their <code>MARGIN</code> argument (<code>apply</code> has
had this for almost a decade).
</p>
</li>
<li><p> New function <code>proportions()</code> and
<code>marginSums()</code>. These should replace the unfortunately named
<code>prop.table()</code> and <code>margin.table()</code>. They are drop-in
replacements, but also add named-margin functionality. The old
function names are retained as aliases for back-compatibility.
</p>
</li>
<li><p> Functions <code>rbinom()</code>, <code>rgeom()</code>, <code>rhyper()</code>,
<code>rpois()</code>, <code>rnbinom(),</code> <code>rsignrank()</code> and
<code>rwilcox()</code> which have returned integer since <span class="rlang"><b>R</b></span> 3.0.0 and hence
<code>NA</code> when the numbers would have been outside the integer range,
now return double vectors (without NAs, typically) in these cases.
</p>
</li>
<li> <p><code>matplot(x,y)</code> (and hence <code>matlines()</code> and
<code>matpoints()</code>) now call the corresponding methods of
<code>plot()</code> and <code>lines()</code>, e.g, when <code>x</code> is a
<code>"Date"</code> or <code>"POSIXct"</code> object; prompted by Spencer Graves'
suggestion.
</p>
</li>
<li> <p><code>stopifnot()</code> now allows customizing error messages via
argument names, thanks to a patch proposal by Neal Fultz in <a href="https://bugs.R-project.org/show_bug.cgi?id=17688">PR#17688</a>.
</p>
</li>
<li> <p><code>unlink()</code> gains a new argument <code>expand</code> to disable
wildcard and tilde expansion. Elements of <code>x</code> of value
<code>"~"</code> are now ignored.
</p>
</li>
<li> <p><code>mle()</code> in the <span class="pkg">stats4</span> package has had its
interface extended so that arguments to the negative
log-likelihood function can be one or more vectors, with similar
conventions applying to bounds, start values, and parameter values
to be kept fixed. This required a minor extension to class
<code>"mle"</code>, so saved objects from earlier versions may need to
be recomputed.
</p>
</li>
<li><p> The default for <code>pdf()</code> is now
<code>useDingbats = FALSE</code>.
</p>
</li>
<li><p> The default fill colour for <code>hist()</code> and
<code>boxplot()</code> is now <code>col = "lightgray"</code>.
</p>
</li>
<li><p> The default order of the levels on the y-axis for
<code>spineplot()</code> and <code>cdplot()</code> has been reversed.
</p>
</li>
<li><p> If the <span class="env">R_ALWAYS_INSTALL_TESTS</span> environment variable is
set to a true value, <code>R CMD INSTALL</code> behaves as if
the <span class="option">--install-tests</span> option is always specified. Thanks to
Reinhold Koch for the suggestion.
</p>
</li>
<li><p> New function <code>R_user_dir()</code> in package <span class="pkg">tools</span>
suggests paths appropriate for storing <span class="rlang"><b>R</b></span>-related user-specific
data, configuration and cache files.
</p>
</li>
<li> <p><code>capabilities()</code> gains a new logical option <code>Xchk</code>
to avoid warnings about X11-related capabilities.
</p>
</li>
<li><p> The internal implementation of <span class="pkg">grid</span> units has changed,
but the only visible effects at user-level should be
</p>
<ul>
<li><p> a slightly different print format for some units
(especially unit arithmetic),
</p>
</li>
<li><p> faster performance (for unit operations) and
</p>
</li>
<li><p> two new functions <code>unitType()</code> and <code>unit.psum()</code>.
</p>
</li></ul>
<p>Based on code contributed by Thomas Lin Pedersen.
</p>
</li>
<li><p> When internal dispatch for <code>rep.int()</code>
and <code>rep_len()</code> fails, there is an attempt to dispatch on the
equivalent call to <code>rep()</code>.
</p>
</li>
<li><p> Object <code>.Machine</code> now contains new <code>longdouble.*</code>
entries (when <span class="rlang"><b>R</b></span> uses long doubles internally).
</p>
</li>
<li> <p><code>news()</code> has been enhanced to cover the news on <span class="rlang"><b>R</b></span> 3.x
and 2.x.
</p>
</li>
<li><p> For consistency, <code>N <- NULL; N[[1]] <- val</code> now turns
<code>N</code> into a <code>list</code> also when <code>val)</code> has length one.
This enables <code>dimnames(r1)[[1]] <- "R1"</code> for a 1-row matrix
<code>r1</code>, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17719">PR#17719</a> reported by Serguei Sokol.
</p>
</li>
<li> <p><code>deparse(..)</code>, <code>dump(..)</code>, and <code>dput(x,
control = "all")</code> now include control option <code>"digits17"</code>
which typically ensures 1:1 invertibility. New option
<code>control = "exact"</code> ensures numeric exact invertibility
via <code>"hexNumeric"</code>.
</p>
</li>
<li><p> When loading data sets via <code>read.table()</code>,
<code>data()</code> now uses ‘<span class="samp">⁠LC_COLLATE=C⁠</span>’ to ensure
locale-independent results for possible string-to-factor
conversions.
</p>
</li>
<li><p> A server socket connection, a new connection type
representing a listening server socket, is created via
<code>serverSocket()</code> and can accept multiple socket connections
via <code>socketAccept()</code>.
</p>
</li>
<li><p> New function <code>socketTimeout()</code> changes the connection
timeout of a socket connection.
</p>
</li>
<li><p> The time needed to start a homogeneous ‘<span class="samp">⁠PSOCK⁠</span>’ cluster on
‘<span class="samp">⁠localhost⁠</span>’ with many nodes has been significantly reduced
(package <span class="pkg">parallel</span>).
</p>
</li>
<li><p> New <code>globalCallingHandlers()</code> function to establish
global condition handlers. This allows registering default
handlers for specific condition classes. Developed in
collaboration with Lionel Henry.
</p>
</li>
<li><p> New function <code>tryInvokeRestart()</code> to invoke a specified
restart if one is available and return without signaling an error
if no such restart is found. Contributed by Lionel Henry in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17598">PR#17598</a>.
</p>
</li>
<li> <p><code>str(x)</code> now shows the length of <code>attributes</code> in some
cases for a data frame <code>x</code>.
</p>
</li>
<li> <p><code>Rprof()</code> gains a new argument <code>filter.callframes</code>
to request that intervening call frames due to lazy evaluation or
explicit <code>eval()</code> calls be omitted from the recorded profile
data. Contributed by Lionel Henry in <a href="https://bugs.R-project.org/show_bug.cgi?id=17595">PR#17595</a>.
</p>
</li>
<li><p> The handling of <code>${FOO-bar}</code> and <code>${FOO:-bar}</code>
in ‘<span class="file">Renviron</span>’ files now follows POSIX shells (at least on a
Unix-alike), so the first treats empty environment variables as
set and the second does not. Previously both ignored empty
variables. There are several uses of the first form in
‘<span class="file">etc/Renviron</span>’.
</p>
</li>
<li><p> New <code>classes</code> argument for <code>suppressWarnings()</code>
and <code>suppressMessages()</code> to selectively suppress only
warnings or messages that inherit from particular classes.
Based on patch from Lionel Henry submitted with <a href="https://bugs.R-project.org/show_bug.cgi?id=17619">PR#17619</a>.
</p>
</li>
<li><p> New function <code>activeBindingFunction()</code> retrieves the
function of an active binding.
</p>
</li>
<li><p> New <code>"cairoFT"</code> and <code>"pango"</code> components in the
output of <code>grSoftVersion()</code>.
</p>
</li>
<li><p> New argument <code>symbolfamily</code> in cairo-based graphics
devices and new function <code>cairoSymbolFont()</code> that can be used
to provide the value for that argument.
</p>
</li></ul>
<h4>Windows</h4>
<ul>
<li> <p><code>Rterm</code> now works also when invoked from <abbr>MSYS2</abbr>
terminals. Line editing is possible when command <code>winpty</code> is
installed.
</p>
</li>
<li> <p><code>normalizePath()</code> now resolves symbolic links and
normalizes case of long names of path elements in case-insensitive
folders (<a href="https://bugs.R-project.org/show_bug.cgi?id=17165">PR#17165</a>).
</p>
</li>
<li> <p><code>md5sum()</code> supports UTF-8 file names with characters
that cannot be translated to the native encoding (<a href="https://bugs.R-project.org/show_bug.cgi?id=17633">PR#17633</a>).
</p>
</li>
<li> <p><code>Rterm</code> gains a new option <span class="option">--workspace</span> to
specify the workspace to be restored. This allows equals to be
part of the name when opening <em>via</em> Windows file associations
(reported by Christian Asseburg).
</p>
</li>
<li> <p><code>Rterm</code> now accepts <code>ALT+xxx</code> sequences also
with NumLock on. Tilde can be pasted with an Italian keyboard
(<a href="https://bugs.R-project.org/show_bug.cgi?id=17679">PR#17679</a>).
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> falls back to copying when junction creation fails during
package checking (patch from Duncan Murdoch).
</p>
</li></ul>
<h4>DEPRECATED AND DEFUNCT</h4>
<ul>
<li><p> Make macro ‘<span class="samp">⁠F77_VISIBILITY⁠</span>’ has been removed
and replaced by ‘<span class="samp">⁠F_VISIBILITY⁠</span>’.
</p>
</li>
<li><p> Make macros ‘<span class="samp">⁠F77⁠</span>’, ‘<span class="samp">⁠FCPIFCPLAGS⁠</span>’ and
‘<span class="samp">⁠SHLIB_OPENMP_FCFLAGS⁠</span>’ have been removed and replaced by
‘<span class="samp">⁠FC⁠</span>’, ‘<span class="samp">⁠FPICFLAGS⁠</span>’ and ‘<span class="samp">⁠SHLIB_OPENMP_FFLAGS⁠</span>’
respectively. (Most <code>make</code> programs will set ‘<span class="samp">⁠F77⁠</span>’
to the value of ‘<span class="samp">⁠FC⁠</span>’, which is set for package compilation.
But portable code should not rely on this.)
</p>
</li>
<li><p> The deprecated support for specifying C++98 for package
installation has been removed.
</p>
</li>
<li> <p><code>R CMD config</code> no longer knows about the
unused settings ‘<span class="samp">⁠F77⁠</span>’ and ‘<span class="samp">⁠FCPIFCPLAGS⁠</span>’, nor
‘<span class="samp">⁠CXX98⁠</span>’ and similar.
</p>
</li>
<li><p> Either PCRE2 or PCRE1 >= 8.32 (Nov 2012) is required: the
deprecated provision for 8.20–8.31 has been removed.
</p>
</li>
<li><p> Defunct functions <code>mem.limits()</code>, <code>.readRDS()</code>,
<code>.saveRDS()</code>, <code>.find.package()</code>, and <code>.path.package()</code>
from package <span class="pkg">base</span>
and <code>allGenerics()</code>, <code>getAccess()</code>, <code>getAllMethods()</code>,
<code>getClassName()</code>, <code>getClassPackage()</code>, <code>getExtends()</code>,
<code>getProperties()</code>, <code>getPrototype()</code>,
<code>getSubclasses()</code>, <code>getVirtual()</code>, <code>mlistMetaName()</code>,
<code>removeMethodsObject()</code>, <code>seemsS4Object()</code>,
<code>traceOff()</code>, and <code>traceOn()</code> from <span class="pkg">methods</span> have been
removed.
</p>
</li></ul>
<h4>C-LEVEL FACILITIES</h4>
<ul>
<li> <p><code>installChar</code> is now remapped in ‘<span class="file">Rinternals.h</span>’ to
<code>installTrChar</code>, of which it has been a wrapper since
<span class="rlang"><b>R</b></span> 3.6.0. Neither are part of the API, but packages using
<code>installChar</code> can replace it if they depend on ‘<span class="samp">⁠R >= 3.6.2⁠</span>’.
</p>
</li>
<li><p> Header ‘<span class="file">R_ext/Print.h</span>’ defines ‘<span class="samp">⁠R_USE_C99_IN_CXX⁠</span>’
and hence exposes <code>Rvprintf</code> and <code>REvprintf</code> if used
with a C++11 (or later) compiler.
</p>
</li>
<li><p> There are new Fortran subroutines <code>dblepr1</code>,
<code>realpr1</code> and <code>intpr1</code> to print a scalar variable
(<code>gfortran</code> 10 enforces the distinction between scalars
and length-one arrays). Also <code>labelpr</code> to print just a label.
</p>
</li>
<li> <p><code>R_withCallingErrorHandler</code> is now available for
establishing a calling handler in C code for conditions inheriting
from class <code>error</code>.
</p>
</li></ul>
<h4>INSTALLATION on a UNIX-ALIKE</h4>
<ul>
<li><p> User-set ‘<span class="samp">⁠DEFS⁠</span>’ (e.g., in ‘<span class="file">config.site</span>’) is now
used for compiling packages (including base packages).
</p>
</li>
<li><p> There is a new variant option <span class="option">--enable-lto=check</span>
for checking consistency of BLAS/LAPACK/LINPACK calls — see
‘Writing R Extensions’.
</p>
</li>
<li><p> A C++ compiler default is set only if the C++11 standard is
supported: it no longer falls back to C++98.
</p>
</li>
<li><p> PCRE2 is used if available. To make use of PCRE1 if PCRE2
is unavailable, configure with option <span class="option">--with-pcre1</span>.
</p>
</li>
<li><p> The minimum required version of <code>libcurl</code> is now 7.28.0
(Oct 2012).
</p>
</li>
<li><p> New make target <code>distcheck</code> checks
</p>
<ul>
<li> <p><span class="rlang"><b>R</b></span> can be rebuilt from the tarball created by
<code>make dist</code>,
</p>
</li>
<li><p> the build from the tarball passes <code>make check-all</code>,
</p>
</li>
<li><p> the build installs and uninstalls,
</p>
</li>
<li><p> the source files are properly cleaned by <code>make distclean</code>.
</p>
</li></ul>
</li></ul>
<h4>UTILITIES</h4>
<ul>
<li> <p><code>R --help</code> now mentions the option <code>--no-echo</code>
(renamed from <code>--slave</code>) and its previously undocumented
short form <code>-s</code>.
</p>
</li>
<li> <p><code>R CMD check</code> now optionally checks
<code>configure</code> and <code>cleanup</code> scripts for
non-Bourne-shell code (‘bashisms’).
</p>
</li>
<li> <p><code>R CMD check --as-cran</code> now runs <code style="white-space: pre;">⁠\donttest⁠</code>
examples (which are run by <code>example()</code>) instead of
instructing the tester to do so. This can be temporarily
circumvented during development by setting environment variable
<span class="env">_R_CHECK_DONTTEST_EXAMPLES_</span> to a false value.
</p>
</li></ul>
<h4>PACKAGE INSTALLATION</h4>
<ul>
<li><p> There is the beginnings of support for the recently approved
C++20 standard, specified analogously to C++14 and C++17. There is
currently only limited support for this in compilers, with flags
such as <span class="option">-std=c++20</span> and <span class="option">-std=c++2a</span>. For the time
being the <code>configure</code> test is of accepting one of these
flags and compiling C++17 code.
</p>
</li></ul>
<h4>BUG FIXES</h4>
<ul>
<li> <p><code>formula(x)</code> with <code>length(x) > 1</code> character vectors,
is deprecated now. Such use has been rare, and has ‘worked’
as expected in some cases only. In other cases, wrong <code>x</code> have
silently been truncated, not detecting previous errors.
</p>
</li>
<li><p> Long-standing issue where the X11 device could lose events
shortly after startup has been addressed (<a href="https://bugs.R-project.org/show_bug.cgi?id=16702">PR#16702</a>).
</p>
</li>
<li><p> The <code>data.frame</code> method for <code>rbind()</code> no longer
drops <code><NA></code> levels from factor columns by default
(<a href="https://bugs.R-project.org/show_bug.cgi?id=17562">PR#17562</a>).
</p>
</li>
<li> <p><code>available.packages()</code> and hence <code>install.packages()</code>
now pass their <code>...</code> argument to <code>download.file()</code>,
fulfilling the wish of <a href="https://bugs.R-project.org/show_bug.cgi?id=17532">PR#17532</a>; subsequently,
<code>available.packages()</code> gets new argument <code>quiet</code>, solving
<a href="https://bugs.R-project.org/show_bug.cgi?id=17573">PR#17573</a>.
</p>
</li>
<li> <p><code>stopifnot()</code> gets new argument <code>exprObject</code> to allow
an <span class="rlang"><b>R</b></span> object of class <code>expression</code> (or other ‘language’)
to work more consistently, thanks to suggestions by Suharto Anggono.
</p>
</li>
<li> <p><code>conformMethod()</code> now works correctly in cases containing
a “<code>&&</code> logic” bug, reported by Henrik Bengtsson. It now
creates methods with <code>"missing"</code> entries in the signature.
Consequently, <code>rematchDefinition()</code> is amended to use
appropriate <code>.local()</code> calls with named arguments where needed.
</p>
</li>
<li> <p><code>format.default(*, scientific = FALSE)</code> now corresponds to
a practically most extreme <code>options(scipen = n)</code> setting rather
than arbitrary <code>n = 100</code>.
</p>
</li>
<li> <p><code>format(as.symbol("foo"))</code> now works (returning <code>"foo"</code>).
</p>
</li>
<li> <p><code>postscript(.., title = *)</code> now signals an error when
the title string contains a character which would produce corrupt
PostScript, thanks to <a href="https://bugs.R-project.org/show_bug.cgi?id=17607">PR#17607</a> by Daisuko Ogawa.
</p>
</li>
<li><p> Certain <code>Ops</code> (notably comparison such as <code>==</code>) now
also work for 0-length data frames, after reports by Hilmar Berger.
</p>
</li>
<li> <p><code>methods(class = class(glm(..)))</code> now warns more usefully
and only once.
</p>
</li>
<li> <p><code>write.dcf()</code> no longer mangles field names (<a href="https://bugs.R-project.org/show_bug.cgi?id=17589">PR#17589</a>).
</p>
</li>
<li><p> Primitive replacement functions no longer mutate a
referenced first argument when used outside of a complex
assignment context.
</p>
</li>
<li><p> A better error message for <code>contour(*, levels = Inf)</code>.
</p>
</li>
<li><p> The return value of <code>contourLines()</code> is no longer
<code>invisible()</code>.
</p>
</li>
<li><p> The Fortran code for calculating the <code>coefficients</code>
component in <code>lm.influence()</code> was very inefficient. It has
(for now) been replaced with much faster <span class="rlang"><b>R</b></span> code (<a href="https://bugs.R-project.org/show_bug.cgi?id=17624">PR#17624</a>).
</p>
</li>
<li> <p><code>cm.colors(n)</code> <em>etc</em> no longer append the code for
<code>alpha = 1</code>, <code>"FF"</code>, to all colors. Hence all eight
<code>*.colors()</code> functions and <code>rainbow()</code> behave
consistently and have the same non-explicit default (<a href="https://bugs.R-project.org/show_bug.cgi?id=17659">PR#17659</a>).
</p>
</li>
<li> <p><code>dnorm</code> had a problematic corner case with <code>sd ==
-Inf</code> or negative <code>sd</code> which was not flagged as an error in
all cases. Thanks to Stephen D. Weigand for reporting and
Wang Jiefei for analyzing this; similar change has been made in
<code>dlnorm()</code>.
</p>
</li>
<li><p> The optional <code>iter.smooth</code> argument of
<code>plot.lm()</code>, (the <code>plot()</code> method for <code>lm</code> and
<code>glm</code> fits) now defaults to <code>0</code> for all <code>glm</code> fits.
Especially for binary observations with high or low fitted
probabilities, this effectively deleted all observations of 1 or
0. Also, the type of residuals used in the <code>glm</code> case has
been switched to <code>"pearson"</code> since deviance residuals do not
in general have approximately zero mean.
</p>
</li>
<li><p> In <code>plot.lm</code>, Cook's distance was computed from unweighted
residuals, leading to inconsistencies. Replaced with usual weighted
version. (<a href="https://bugs.R-project.org/show_bug.cgi?id=16056">PR#16056</a>)
</p>
</li>
<li><p> Time-series <code>ts(*, start, end, frequency)</code> with
fractional <code>frequency</code> are supported more consistently;
thanks to a report from Johann Kleinbub and analysis and patch by
Duncan Murdoch in <a href="https://bugs.R-project.org/show_bug.cgi?id=17669">PR#17669</a>.
</p>
</li>
<li><p> In case of errors <code>mcmapply()</code> now preserves attributes
of returned <code>"try-error"</code> objects and avoids simplification,
overriding <code>SIMPLIFY</code> to <code>FALSE</code>. (<a href="https://bugs.R-project.org/show_bug.cgi?id=17653">PR#17653</a>)
</p>
</li>
<li> <p><code>as.difftime()</code> gets new optional <code>tz = "UTC"</code>
argument which should fix behaviour during
daylight-savings-changeover days, fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=16764">PR#16764</a>, thanks to
proposals and analysis by Johannes Ranke and Kirill Müller.
</p>
</li>
<li> <p><code>round()</code> does a better job of rounding
<em>“to nearest”</em> by <em>measuring</em> and
<em>“to even”</em>; thanks to a careful algorithm originally
prompted by the report from Adam Wheeler and then others, in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17668">PR#17668</a>. <br> <code>round(x, dig)</code> for <em>negative</em> digits
is much more rational now, notably for large <code class="reqn">|dig|</code>.
</p>
</li>
<li><p> Inheritance information on S4 classes is maintained more
consistently, particularly in the case of class unions (in part
due to <a href="https://bugs.R-project.org/show_bug.cgi?id=17596">PR#17596</a> and a report from Ezra Tucker).
</p>
</li>
<li> <p><code>is()</code> behaves more robustly when its argument
<code>class2</code> is a <code>classRepresentation</code> object.
</p>
</li>
<li><p> The warning message when attempting to export an nonexistent
class is now more readable; thanks to Thierry Onkelinx for
recognizing the problem.
</p>
</li>
<li> <p><code>choose()</code> misbehaved in corner cases where it switched
<code>n - k</code> for <code>k</code> and <code>n</code> was only <em>nearly</em> integer
(report from Erik Scott Wright).
</p>
</li>
<li> <p><code>mle()</code> in the <span class="pkg">stats4</span> package had problems
combining use of box constraints and fixed starting values (in
particular, confidence intervals were affected).
</p>
</li>
<li><p> Operator <code>?</code> now has lower precedence than <code>=</code> to
work as documented, so <code>=</code> behaves like <code><-</code> in help
expressions (<a href="https://bugs.R-project.org/show_bug.cgi?id=16710">PR#16710</a>).
</p>
</li>
<li> <p><code>smoothEnds(x)</code> now returns <code>integer</code> type in
<em>both</em> cases when <code>x</code> is <code>integer</code>, thanks to a
report and proposal by Bill Dunlap <a href="https://bugs.R-project.org/show_bug.cgi?id=17693">PR#17693</a>.
</p>
</li>
<li><p> The <span class="pkg">methods</span> package does a better job of tracking
inheritance relationships across packages.
</p>
</li>
<li> <p><code>norm(diag(c(1, NA)), "2")</code> now works.
</p>
</li>
<li> <p><code>subset()</code> had problems with 0-col dataframes (reported
by Bill Dunlap, <a href="https://bugs.R-project.org/show_bug.cgi?id=17721">PR#17721</a>).
</p>
</li>
<li><p> Several cases of integer overflow detected by the
‘undefined behaviour sanitizer’ of <code>clang</code> 10 have
been circumvented. One in <code>rhyper()</code> may change the
generated value for large input values.
</p>
</li>
<li> <p><code>dotchart()</code> now places the y-axis label (<code>ylab</code>)
much better, not overplotting labels, thanks to a report and
suggestion by Alexey Shipunov.
</p>
</li>
<li><p> A rare C-level array overflow in <code>chull()</code> has been
worked around.
</p>
</li>
<li><p> Some invalid specifications of the day-of-the-year
(<em>via</em> <code>%j</code>, e.g. day 366 in 2017) or week
plus day-of-the-week are now detected by <code>strptime()</code>.
They now return <code>NA</code> but give a warning as they may have
given random results or corrupted memory in earlier versions of <span class="rlang"><b>R</b></span>.
</p>
</li>
<li> <p><code>socketConnection(server = FALSE)</code> now respects the
connection timeout also on Linux.
</p>
</li>
<li> <p><code>socketConnection(server = FALSE)</code> no longer leaks a
connection that is available right away without waiting (e.g. on
‘<span class="samp">⁠localhost⁠</span>’).
</p>
</li>
<li><p> Socket connections are now robust against spurious readability
and spurious availability of an incoming connection.
</p>
</li>
<li> <p><code>blocking = FALSE</code> is now respected also on the server side
of a socket connection, allowing non-blocking read operations.
</p>
</li>
<li> <p><code>anova.glm()</code> and <code>anova.glmlist()</code> computed
incorrect score (Rao) tests in no-intercept cases. (André
Gillibert, <a href="https://bugs.R-project.org/show_bug.cgi?id=17735">PR#17735</a>)
</p>
</li>
<li> <p><code>summaryRprof()</code> now should work correctly for the
<code>Rprof(*, memory.profiling=TRUE)</code> case with small chunk size (and
<code>"tseries"</code> or similar) thanks to a patch proposal by
Benjamin Tyner, in <a href="https://bugs.R-project.org/show_bug.cgi?id=15886">PR#15886</a>.
</p>
</li>
<li> <p><code>xgettext()</code> ignores strings passed to
<code>ngettext()</code>, since the latter is handled by
<code>xngettext()</code>. Thanks to Daniele Medri for the report and all
the recent work he has done on the Italian translations.
</p>
</li>
<li> <p><code>data(package = "P")</code> for <code>P</code> in <span class="pkg">base</span> and
<span class="pkg">stats</span> no longer reports the data sets from package
<span class="pkg">datasets</span> (which it did for back compatibility for 16 years),
fixing <a href="https://bugs.R-project.org/show_bug.cgi?id=17730">PR#17730</a>.
</p>
</li>
<li> <p><code>x[[Inf]]</code> (returning <code>NULL</code>) no longer leads to
undefined behavior, thanks to a report by Kirill Müller in
<a href="https://bugs.R-project.org/show_bug.cgi?id=17756">PR#17756</a>. Further, <code>x[[-Inf]]</code> and <code>x[[-n]]</code> now give
more helpful error messages.
</p>
</li>
<li> <p><code>Gamma()</code> family sometimes had trouble storing
link name <a href="https://bugs.R-project.org/show_bug.cgi?id=15891">PR#15891</a>
</p>
</li></ul>
<h4>BUG FIXES (Windows)</h4>
<ul>
<li> <p><code>Sys.glob()</code> now supports all characters
from the Unicode Basic Multilingual Plane, no longer corrupting
some (less commonly used) characters (<a href="https://bugs.R-project.org/show_bug.cgi?id=17638">PR#17638</a>).
</p>
</li>
<li> <p><code>Rterm</code> now correctly displays
multi-byte-coded characters representable in the current native
encoding (at least on Windows 10 they were sometimes omitted,
<a href="https://bugs.R-project.org/show_bug.cgi?id=17632">PR#17632</a>).
</p>
</li>
<li> <p><code>scan()</code> issues with UTF-8 data when running in a <abbr>DBCS</abbr>
locale have been resolved (<a href="https://bugs.R-project.org/show_bug.cgi?id=16520">PR#16520</a>, <a href="https://bugs.R-project.org/show_bug.cgi?id=16584">PR#16584</a>).
</p>
</li>
<li> <p><code>Rterm</code> now accepts enhanced/arrow keys also with
ConPTY.
</p>
</li>
<li> <p><span class="rlang"><b>R</b></span> can now be started <em>via</em> the launcher icon in a
user documents directory whose path is not representable in the
system encoding.
</p>
</li>
<li> <p><code>socketConnection(server = FALSE)</code> now returns instantly
also on Windows when connection failure is signalled.
</p>
</li>
<li><p> Problems with UTF-16 surrogate pairs have been fixed in
several functions, including <code>tolower()</code> and
<code>toupper()</code> (<a href="https://bugs.R-project.org/show_bug.cgi?id=17645">PR#17645</a>).
</p>
</li></ul>
<h3>CHANGES in previous versions</h3>
<ul>
<li><p> Older news can be found in text format in files
<a href="../NEWS.0">NEWS.0</a>, <a href="../NEWS.1">NEWS.1</a>,
<a href="../NEWS.2">NEWS.2</a> and
<a href="../NEWS.3">NEWS.3</a>
in the ‘<span class="file">doc</span>’ directory. News in HTML format for
<span class="rlang"><b>R</b></span> versions 3.x and from 2.10.0 to 2.15.3 are available at
<a href="NEWS.3.html">NEWS.3.html</a>
and
<a href="NEWS.2.html">NEWS.2.html</a>.
</p>
</li></ul>
</main>
</div>
</body></html>
|
