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
|
#! /usr/bin/env python
## -*- Mode: python; py-indent-offset: 4; indent-tabs-mode: nil; coding: utf-8; -*-
#
# Copyright (c) 2009 University of Washington
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation;
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
#
from __future__ import print_function
import os
import sys
import time
import optparse
import subprocess
import threading
import signal
import xml.dom.minidom
import shutil
import re
from utils import get_list_from_file
try:
import queue
except ImportError:
import Queue as queue
#
# XXX This should really be part of a waf command to list the configuration
# items relative to optional ns-3 pieces.
#
# A list of interesting configuration items in the waf configuration
# cache which we may be interested in when deciding on which examples
# to run and how to run them. These are set by waf during the
# configuration phase and the corresponding assignments are usually
# found in the associated subdirectory wscript files.
#
interesting_config_items = [
"NS3_ENABLED_MODULES",
"NS3_MODULE_PATH",
"NSC_ENABLED",
"ENABLE_REAL_TIME",
"ENABLE_THREADING",
"ENABLE_EXAMPLES",
"ENABLE_TESTS",
"EXAMPLE_DIRECTORIES",
"ENABLE_PYTHON_BINDINGS",
"NSCLICK",
"ENABLE_BRITE",
"ENABLE_OPENFLOW",
"APPNAME",
"BUILD_PROFILE",
"VERSION",
"PYTHON",
"VALGRIND_FOUND",
]
NSC_ENABLED = False
ENABLE_REAL_TIME = False
ENABLE_THREADING = False
ENABLE_EXAMPLES = True
ENABLE_TESTS = True
NSCLICK = False
ENABLE_BRITE = False
ENABLE_OPENFLOW = False
EXAMPLE_DIRECTORIES = []
APPNAME = ""
BUILD_PROFILE = ""
BUILD_PROFILE_SUFFIX = ""
VERSION = ""
PYTHON = ""
VALGRIND_FOUND = True
#
# This will be given a prefix and a suffix when the waf config file is
# read.
#
test_runner_name = "test-runner"
#
# If the user has constrained us to run certain kinds of tests, we can tell waf
# to only build
#
core_kinds = ["bvt", "core", "performance", "system", "unit"]
#
# There are some special cases for test suites that kill valgrind. This is
# because NSC causes illegal instruction crashes when run under valgrind.
#
core_valgrind_skip_tests = [
"ns3-tcp-cwnd",
"nsc-tcp-loss",
"ns3-tcp-interoperability",
"routing-click",
"lte-rr-ff-mac-scheduler",
"lte-tdmt-ff-mac-scheduler",
"lte-fdmt-ff-mac-scheduler",
"lte-pf-ff-mac-scheduler",
"lte-tta-ff-mac-scheduler",
"lte-fdbet-ff-mac-scheduler",
"lte-ttbet-ff-mac-scheduler",
"lte-fdtbfq-ff-mac-scheduler",
"lte-tdtbfq-ff-mac-scheduler",
"lte-pss-ff-mac-scheduler",
]
#
# There are some special cases for test suites that fail when NSC is
# missing.
#
core_nsc_missing_skip_tests = [
"ns3-tcp-cwnd",
"nsc-tcp-loss",
"ns3-tcp-interoperability",
]
#
# Parse the examples-to-run file if it exists.
#
# This function adds any C++ examples or Python examples that are to be run
# to the lists in example_tests and python_tests, respectively.
#
def parse_examples_to_run_file(
examples_to_run_path,
cpp_executable_dir,
python_script_dir,
example_tests,
example_names_original,
python_tests):
# Look for the examples-to-run file exists.
if os.path.exists(examples_to_run_path):
# Each tuple in the C++ list of examples to run contains
#
# (example_name, do_run, do_valgrind_run)
#
# where example_name is the executable to be run, do_run is a
# condition under which to run the example, and do_valgrind_run is
# a condition under which to run the example under valgrind. This
# is needed because NSC causes illegal instruction crashes with
# some tests when they are run under valgrind.
#
# Note that the two conditions are Python statements that
# can depend on waf configuration variables. For example,
#
# ("tcp-nsc-lfn", "NSC_ENABLED == True", "NSC_ENABLED == False"),
#
cpp_examples = get_list_from_file(examples_to_run_path, "cpp_examples")
for example_name, do_run, do_valgrind_run in cpp_examples:
# Seperate the example name from its arguments.
example_name_original = example_name
example_name_parts = example_name.split(' ', 1)
if len(example_name_parts) == 1:
example_name = example_name_parts[0]
example_arguments = ""
else:
example_name = example_name_parts[0]
example_arguments = example_name_parts[1]
# Add the proper prefix and suffix to the example name to
# match what is done in the wscript file.
example_path = "%s%s-%s%s" % (APPNAME, VERSION, example_name, BUILD_PROFILE_SUFFIX)
# Set the full path for the example.
example_path = os.path.join(cpp_executable_dir, example_path)
example_name = os.path.join(
os.path.relpath(cpp_executable_dir, NS3_BUILDDIR),
example_name)
# Add all of the C++ examples that were built, i.e. found
# in the directory, to the list of C++ examples to run.
if os.path.exists(example_path):
# Add any arguments to the path.
if len(example_name_parts) != 1:
example_path = "%s %s" % (example_path, example_arguments)
example_name = "%s %s" % (example_name, example_arguments)
# Add this example.
example_tests.append((example_name, example_path, do_run, do_valgrind_run))
example_names_original.append(example_name_original)
# Each tuple in the Python list of examples to run contains
#
# (example_name, do_run)
#
# where example_name is the Python script to be run and
# do_run is a condition under which to run the example.
#
# Note that the condition is a Python statement that can
# depend on waf configuration variables. For example,
#
# ("realtime-udp-echo.py", "ENABLE_REAL_TIME == True"),
#
python_examples = get_list_from_file(examples_to_run_path, "python_examples")
for example_name, do_run in python_examples:
# Seperate the example name from its arguments.
example_name_parts = example_name.split(' ', 1)
if len(example_name_parts) == 1:
example_name = example_name_parts[0]
example_arguments = ""
else:
example_name = example_name_parts[0]
example_arguments = example_name_parts[1]
# Set the full path for the example.
example_path = os.path.join(python_script_dir, example_name)
# Add all of the Python examples that were found to the
# list of Python examples to run.
if os.path.exists(example_path):
# Add any arguments to the path.
if len(example_name_parts) != 1:
example_path = "%s %s" % (example_path, example_arguments)
# Add this example.
python_tests.append((example_path, do_run))
#
# The test suites are going to want to output status. They are running
# concurrently. This means that unless we are careful, the output of
# the test suites will be interleaved. Rather than introducing a lock
# file that could unintentionally start serializing execution, we ask
# the tests to write their output to a temporary directory and then
# put together the final output file when we "join" the test tasks back
# to the main thread. In addition to this issue, the example programs
# often write lots and lots of trace files which we will just ignore.
# We put all of them into the temp directory as well, so they can be
# easily deleted.
#
TMP_OUTPUT_DIR = "testpy-output"
def read_test(test):
result = test.find('Result').text
name = test.find('Name').text
if not test.find('Time') is None:
time_real = test.find('Time').get('real')
else:
time_real = ''
return (result, name, time_real)
#
# A simple example of writing a text file with a test result summary. It is
# expected that this output will be fine for developers looking for problems.
#
def node_to_text (test, f):
(result, name, time_real) = read_test(test)
output = "%s: Test Suite \"%s\" (%s)\n" % (result, name, time_real)
f.write(output)
for details in test.findall('FailureDetails'):
f.write(" Details:\n")
f.write(" Message: %s\n" % details.find('Message').text)
f.write(" Condition: %s\n" % details.find('Condition').text)
f.write(" Actual: %s\n" % details.find('Actual').text)
f.write(" Limit: %s\n" % details.find('Limit').text)
f.write(" File: %s\n" % details.find('File').text)
f.write(" Line: %s\n" % details.find('Line').text)
for child in test.findall('Test'):
node_to_text(child, f)
def translate_to_text(results_file, text_file):
f = open(text_file, 'w')
import xml.etree.ElementTree as ET
et = ET.parse (results_file)
for test in et.findall('Test'):
node_to_text (test, f)
for example in et.findall('Example'):
result = example.find('Result').text
name = example.find('Name').text
if not example.find('Time') is None:
time_real = example.find('Time').get('real')
else:
time_real = ''
output = "%s: Example \"%s\" (%s)\n" % (result, name, time_real)
f.write(output)
f.close()
#
# A simple example of writing an HTML file with a test result summary. It is
# expected that this will eventually be made prettier as time progresses and
# we have time to tweak it. This may end up being moved to a separate module
# since it will probably grow over time.
#
def translate_to_html(results_file, html_file):
f = open(html_file, 'w')
f.write("<html>\n")
f.write("<body>\n")
f.write("<center><h1>ns-3 Test Results</h1></center>\n")
#
# Read and parse the whole results file.
#
import xml.etree.ElementTree as ET
et = ET.parse(results_file)
#
# Iterate through the test suites
#
f.write("<h2>Test Suites</h2>\n")
for suite in et.findall('Test'):
#
# For each test suite, get its name, result and execution time info
#
(result, name, time) = read_test (suite)
#
# Print a level three header with the result, name and time. If the
# test suite passed, the header is printed in green. If the suite was
# skipped, print it in orange, otherwise assume something bad happened
# and print in red.
#
if result == "PASS":
f.write("<h3 style=\"color:green\">%s: %s (%s)</h3>\n" % (result, name, time))
elif result == "SKIP":
f.write("<h3 style=\"color:#ff6600\">%s: %s (%s)</h3>\n" % (result, name, time))
else:
f.write("<h3 style=\"color:red\">%s: %s (%s)</h3>\n" % (result, name, time))
#
# The test case information goes in a table.
#
f.write("<table border=\"1\">\n")
#
# The first column of the table has the heading Result
#
f.write("<th> Result </th>\n")
#
# If the suite crashed or is skipped, there is no further information, so just
# delare a new table row with the result (CRASH or SKIP) in it. Looks like:
#
# +--------+
# | Result |
# +--------+
# | CRASH |
# +--------+
#
# Then go on to the next test suite. Valgrind and skipped errors look the same.
#
if result in ["CRASH", "SKIP", "VALGR"]:
f.write("<tr>\n")
if result == "SKIP":
f.write("<td style=\"color:#ff6600\">%s</td>\n" % result)
else:
f.write("<td style=\"color:red\">%s</td>\n" % result)
f.write("</tr>\n")
f.write("</table>\n")
continue
#
# If the suite didn't crash, we expect more information, so fill out
# the table heading row. Like,
#
# +--------+----------------+------+
# | Result | Test Case Name | Time |
# +--------+----------------+------+
#
f.write("<th>Test Case Name</th>\n")
f.write("<th> Time </th>\n")
#
# If the test case failed, we need to print out some failure details
# so extend the heading row again. Like,
#
# +--------+----------------+------+-----------------+
# | Result | Test Case Name | Time | Failure Details |
# +--------+----------------+------+-----------------+
#
if result == "FAIL":
f.write("<th>Failure Details</th>\n")
#
# Now iterate through all of the test cases.
#
for case in suite.findall('Test'):
#
# Get the name, result and timing information from xml to use in
# printing table below.
#
(result, name, time) = read_test(case)
#
# If the test case failed, we iterate through possibly multiple
# failure details
#
if result == "FAIL":
#
# There can be multiple failures for each test case. The first
# row always gets the result, name and timing information along
# with the failure details. Remaining failures don't duplicate
# this information but just get blanks for readability. Like,
#
# +--------+----------------+------+-----------------+
# | Result | Test Case Name | Time | Failure Details |
# +--------+----------------+------+-----------------+
# | FAIL | The name | time | It's busted |
# +--------+----------------+------+-----------------+
# | | | | Really broken |
# +--------+----------------+------+-----------------+
# | | | | Busted bad |
# +--------+----------------+------+-----------------+
#
first_row = True
for details in case.findall('FailureDetails'):
#
# Start a new row in the table for each possible Failure Detail
#
f.write("<tr>\n")
if first_row:
first_row = False
f.write("<td style=\"color:red\">%s</td>\n" % result)
f.write("<td>%s</td>\n" % name)
f.write("<td>%s</td>\n" % time)
else:
f.write("<td></td>\n")
f.write("<td></td>\n")
f.write("<td></td>\n")
f.write("<td>")
f.write("<b>Message: </b>%s, " % details.find('Message').text)
f.write("<b>Condition: </b>%s, " % details.find('Condition').text)
f.write("<b>Actual: </b>%s, " % details.find('Actual').text)
f.write("<b>Limit: </b>%s, " % details.find('Limit').text)
f.write("<b>File: </b>%s, " % details.find('File').text)
f.write("<b>Line: </b>%s" % details.find('Line').text)
f.write("</td>\n")
#
# End the table row
#
f.write("</td>\n")
else:
#
# If this particular test case passed, then we just print the PASS
# result in green, followed by the test case name and its execution
# time information. These go off in <td> ... </td> table data.
# The details table entry is left blank.
#
# +--------+----------------+------+---------+
# | Result | Test Case Name | Time | Details |
# +--------+----------------+------+---------+
# | PASS | The name | time | |
# +--------+----------------+------+---------+
#
f.write("<tr>\n")
f.write("<td style=\"color:green\">%s</td>\n" % result)
f.write("<td>%s</td>\n" % name)
f.write("<td>%s</td>\n" % time)
f.write("<td></td>\n")
f.write("</tr>\n")
#
# All of the rows are written, so we need to end the table.
#
f.write("</table>\n")
#
# That's it for all of the test suites. Now we have to do something about
# our examples.
#
f.write("<h2>Examples</h2>\n")
#
# Example status is rendered in a table just like the suites.
#
f.write("<table border=\"1\">\n")
#
# The table headings look like,
#
# +--------+--------------+--------------+
# | Result | Example Name | Elapsed Time |
# +--------+--------------+--------------+
#
f.write("<th> Result </th>\n")
f.write("<th>Example Name</th>\n")
f.write("<th>Elapsed Time</th>\n")
#
# Now iterate through all of the examples
#
for example in et.findall("Example"):
#
# Start a new row for each example
#
f.write("<tr>\n")
#
# Get the result and name of the example in question
#
(result, name, time) = read_test(example)
#
# If the example either failed or crashed, print its result status
# in red; otherwise green. This goes in a <td> ... </td> table data
#
if result == "PASS":
f.write("<td style=\"color:green\">%s</td>\n" % result)
elif result == "SKIP":
f.write("<td style=\"color:#ff6600\">%s</fd>\n" % result)
else:
f.write("<td style=\"color:red\">%s</td>\n" % result)
#
# Write the example name as a new tag data.
#
f.write("<td>%s</td>\n" % name)
#
# Write the elapsed time as a new tag data.
#
f.write("<td>%s</td>\n" % time)
#
# That's it for the current example, so terminate the row.
#
f.write("</tr>\n")
#
# That's it for the table of examples, so terminate the table.
#
f.write("</table>\n")
#
# And that's it for the report, so finish up.
#
f.write("</body>\n")
f.write("</html>\n")
f.close()
#
# Python Control-C handling is broken in the presence of multiple threads.
# Signals get delivered to the runnable/running thread by default and if
# it is blocked, the signal is simply ignored. So we hook sigint and set
# a global variable telling the system to shut down gracefully.
#
thread_exit = False
def sigint_hook(signal, frame):
global thread_exit
thread_exit = True
return 0
#
# In general, the build process itself naturally takes care of figuring out
# which tests are built into the test runner. For example, if waf configure
# determines that ENABLE_EMU is false due to some missing dependency,
# the tests for the emu net device simply will not be built and will
# therefore not be included in the built test runner.
#
# Examples, however, are a different story. In that case, we are just given
# a list of examples that could be run. Instead of just failing, for example,
# nsc-tcp-zoo if NSC is not present, we look into the waf saved configuration
# for relevant configuration items.
#
# XXX This function pokes around in the waf internal state file. To be a
# little less hacky, we should add a commmand to waf to return this info
# and use that result.
#
def read_waf_config():
for line in open(".lock-waf_" + sys.platform + "_build", "rt"):
if line.startswith("top_dir ="):
key, val = line.split('=')
top_dir = eval(val.strip())
if line.startswith("out_dir ="):
key, val = line.split('=')
out_dir = eval(val.strip())
global NS3_BASEDIR
NS3_BASEDIR = top_dir
global NS3_BUILDDIR
NS3_BUILDDIR = out_dir
for line in open("%s/c4che/_cache.py" % out_dir).readlines():
for item in interesting_config_items:
if line.startswith(item):
exec(line, globals())
if options.verbose:
for item in interesting_config_items:
print("%s ==" % item, eval(item))
#
# It seems pointless to fork a process to run waf to fork a process to run
# the test runner, so we just run the test runner directly. The main thing
# that waf would do for us would be to sort out the shared library path but
# we can deal with that easily and do here.
#
# There can be many different ns-3 repositories on a system, and each has
# its own shared libraries, so ns-3 doesn't hardcode a shared library search
# path -- it is cooked up dynamically, so we do that too.
#
def make_paths():
have_DYLD_LIBRARY_PATH = False
have_LD_LIBRARY_PATH = False
have_PATH = False
have_PYTHONPATH = False
keys = list(os.environ.keys())
for key in keys:
if key == "DYLD_LIBRARY_PATH":
have_DYLD_LIBRARY_PATH = True
if key == "LD_LIBRARY_PATH":
have_LD_LIBRARY_PATH = True
if key == "PATH":
have_PATH = True
if key == "PYTHONPATH":
have_PYTHONPATH = True
pypath = os.environ["PYTHONPATH"] = os.path.join (NS3_BUILDDIR, "bindings", "python")
if not have_PYTHONPATH:
os.environ["PYTHONPATH"] = pypath
else:
os.environ["PYTHONPATH"] += ":" + pypath
if options.verbose:
print("os.environ[\"PYTHONPATH\"] == %s" % os.environ["PYTHONPATH"])
if sys.platform == "darwin":
if not have_DYLD_LIBRARY_PATH:
os.environ["DYLD_LIBRARY_PATH"] = ""
for path in NS3_MODULE_PATH:
os.environ["DYLD_LIBRARY_PATH"] += ":" + path
if options.verbose:
print("os.environ[\"DYLD_LIBRARY_PATH\"] == %s" % os.environ["DYLD_LIBRARY_PATH"])
elif sys.platform == "win32":
if not have_PATH:
os.environ["PATH"] = ""
for path in NS3_MODULE_PATH:
os.environ["PATH"] += ';' + path
if options.verbose:
print("os.environ[\"PATH\"] == %s" % os.environ["PATH"])
elif sys.platform == "cygwin":
if not have_PATH:
os.environ["PATH"] = ""
for path in NS3_MODULE_PATH:
os.environ["PATH"] += ":" + path
if options.verbose:
print("os.environ[\"PATH\"] == %s" % os.environ["PATH"])
else:
if not have_LD_LIBRARY_PATH:
os.environ["LD_LIBRARY_PATH"] = ""
for path in NS3_MODULE_PATH:
os.environ["LD_LIBRARY_PATH"] += ":" + str(path)
if options.verbose:
print("os.environ[\"LD_LIBRARY_PATH\"] == %s" % os.environ["LD_LIBRARY_PATH"])
#
# Short note on generating suppressions:
#
# See the valgrind documentation for a description of suppressions. The easiest
# way to generate a suppression expression is by using the valgrind
# --gen-suppressions option. To do that you have to figure out how to run the
# test in question.
#
# If you do "test.py -v -g -s <suitename> then test.py will output most of what
# you need. For example, if you are getting a valgrind error in the
# devices-mesh-dot11s-regression test suite, you can run:
#
# ./test.py -v -g -s devices-mesh-dot11s-regression
#
# You should see in the verbose output something that looks like:
#
# Synchronously execute valgrind --suppressions=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/testpy.supp
# --leak-check=full --error-exitcode=2 /home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/build/debug/utils/ns3-dev-test-runner-debug
# --suite=devices-mesh-dot11s-regression --basedir=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev
# --tempdir=testpy-output/2010-01-12-22-47-50-CUT
# --out=testpy-output/2010-01-12-22-47-50-CUT/devices-mesh-dot11s-regression.xml
#
# You need to pull out the useful pieces, and so could run the following to
# reproduce your error:
#
# valgrind --suppressions=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/testpy.supp
# --leak-check=full --error-exitcode=2 /home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/build/debug/utils/ns3-dev-test-runner-debug
# --suite=devices-mesh-dot11s-regression --basedir=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev
# --tempdir=testpy-output
#
# Hint: Use the first part of the command as is, and point the "tempdir" to
# somewhere real. You don't need to specify an "out" file.
#
# When you run the above command you should see your valgrind error. The
# suppression expression(s) can be generated by adding the --gen-suppressions=yes
# option to valgrind. Use something like:
#
# valgrind --gen-suppressions=yes --suppressions=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/testpy.supp
# --leak-check=full --error-exitcode=2 /home/craigdo/repos/ns-3-allinone-dev/ns-3-dev/build/debug/utils/ns3-dev-test-runner-debug
# --suite=devices-mesh-dot11s-regression --basedir=/home/craigdo/repos/ns-3-allinone-dev/ns-3-dev
# --tempdir=testpy-output
#
# Now when valgrind detects an error it will ask:
#
# ==27235== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ----
#
# to which you just enter 'y'<ret>.
#
# You will be provided with a suppression expression that looks something like
# the following:
# {
# <insert_a_suppression_name_here>
# Memcheck:Addr8
# fun:_ZN3ns36dot11s15HwmpProtocolMac8SendPreqESt6vectorINS0_6IePreqESaIS3_EE
# fun:_ZN3ns36dot11s15HwmpProtocolMac10SendMyPreqEv
# fun:_ZN3ns36dot11s15HwmpProtocolMac18RequestDestinationENS_12Mac48AddressEjj
# ...
# the rest of the stack frame
# ...
# }
#
# You need to add a supression name which will only be printed out by valgrind in
# verbose mode (but it needs to be there in any case). The entire stack frame is
# shown to completely characterize the error, but in most cases you won't need
# all of that info. For example, if you want to turn off all errors that happen
# when the function (fun:) is called, you can just delete the rest of the stack
# frame. You can also use wildcards to make the mangled signatures more readable.
#
# I added the following to the testpy.supp file for this particular error:
#
# {
# Supress invalid read size errors in SendPreq() when using HwmpProtocolMac
# Memcheck:Addr8
# fun:*HwmpProtocolMac*SendPreq*
# }
#
# Now, when you run valgrind the error will be suppressed.
#
VALGRIND_SUPPRESSIONS_FILE = "testpy.supp"
def run_job_synchronously(shell_command, directory, valgrind, is_python, build_path=""):
suppressions_path = os.path.join (NS3_BASEDIR, VALGRIND_SUPPRESSIONS_FILE)
if is_python:
path_cmd = PYTHON[0] + " " + os.path.join (NS3_BASEDIR, shell_command)
else:
if len(build_path):
path_cmd = os.path.join (build_path, shell_command)
else:
path_cmd = os.path.join (NS3_BUILDDIR, shell_command)
if valgrind:
cmd = "valgrind --suppressions=%s --leak-check=full --show-reachable=yes --error-exitcode=2 %s" % (suppressions_path,
path_cmd)
else:
cmd = path_cmd
if options.verbose:
print("Synchronously execute %s" % cmd)
start_time = time.time()
proc = subprocess.Popen(cmd, shell = True, cwd = directory, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout_results, stderr_results = proc.communicate()
elapsed_time = time.time() - start_time
retval = proc.returncode
try:
stdout_results = stdout_results.decode()
except UnicodeDecodeError:
print("Non-decodable character in stdout output of %s" % cmd)
print(stdout_results)
retval = 1
try:
stderr_results = stderr_results.decode()
except UnicodeDecodeError:
print("Non-decodable character in stderr output of %s" % cmd)
print(stderr_results)
retval = 1
#
# valgrind sometimes has its own idea about what kind of memory management
# errors are important. We want to detect *any* leaks, so the way to do
# that is to look for the presence of a valgrind leak summary section.
#
# If another error has occurred (like a test suite has failed), we don't
# want to trump that error, so only do the valgrind output scan if the
# test has otherwise passed (return code was zero).
#
if valgrind and retval == 0 and "== LEAK SUMMARY:" in stderr_results:
retval = 2
if options.verbose:
print("Return code = ", retval)
print("stderr = ", stderr_results)
return (retval, stdout_results, stderr_results, elapsed_time)
#
# This class defines a unit of testing work. It will typically refer to
# a test suite to run using the test-runner, or an example to run directly.
#
class Job:
def __init__(self):
self.is_break = False
self.is_skip = False
self.is_example = False
self.is_pyexample = False
self.shell_command = ""
self.display_name = ""
self.basedir = ""
self.tempdir = ""
self.cwd = ""
self.tmp_file_name = ""
self.returncode = False
self.elapsed_time = 0
self.build_path = ""
#
# A job is either a standard job or a special job indicating that a worker
# thread should exist. This special job is indicated by setting is_break
# to true.
#
def set_is_break(self, is_break):
self.is_break = is_break
#
# If a job is to be skipped, we actually run it through the worker threads
# to keep the PASS, FAIL, CRASH and SKIP processing all in one place.
#
def set_is_skip(self, is_skip):
self.is_skip = is_skip
#
# Examples are treated differently than standard test suites. This is
# mostly because they are completely unaware that they are being run as
# tests. So we have to do some special case processing to make them look
# like tests.
#
def set_is_example(self, is_example):
self.is_example = is_example
#
# Examples are treated differently than standard test suites. This is
# mostly because they are completely unaware that they are being run as
# tests. So we have to do some special case processing to make them look
# like tests.
#
def set_is_pyexample(self, is_pyexample):
self.is_pyexample = is_pyexample
#
# This is the shell command that will be executed in the job. For example,
#
# "utils/ns3-dev-test-runner-debug --test-name=some-test-suite"
#
def set_shell_command(self, shell_command):
self.shell_command = shell_command
#
# This is the build path where ns-3 was built. For example,
#
# "/home/craigdo/repos/ns-3-allinone-test/ns-3-dev/build/debug"
#
def set_build_path(self, build_path):
self.build_path = build_path
#
# This is the dispaly name of the job, typically the test suite or example
# name. For example,
#
# "some-test-suite" or "udp-echo"
#
def set_display_name(self, display_name):
self.display_name = display_name
#
# This is the base directory of the repository out of which the tests are
# being run. It will be used deep down in the testing framework to determine
# where the source directory of the test was, and therefore where to find
# provided test vectors. For example,
#
# "/home/user/repos/ns-3-dev"
#
def set_basedir(self, basedir):
self.basedir = basedir
#
# This is the directory to which a running test suite should write any
# temporary files.
#
def set_tempdir(self, tempdir):
self.tempdir = tempdir
#
# This is the current working directory that will be given to an executing
# test as it is being run. It will be used for examples to tell them where
# to write all of the pcap files that we will be carefully ignoring. For
# example,
#
# "/tmp/unchecked-traces"
#
def set_cwd(self, cwd):
self.cwd = cwd
#
# This is the temporary results file name that will be given to an executing
# test as it is being run. We will be running all of our tests in parallel
# so there must be multiple temporary output files. These will be collected
# into a single XML file at the end and then be deleted.
#
def set_tmp_file_name(self, tmp_file_name):
self.tmp_file_name = tmp_file_name
#
# The return code received when the job process is executed.
#
def set_returncode(self, returncode):
self.returncode = returncode
#
# The elapsed real time for the job execution.
#
def set_elapsed_time(self, elapsed_time):
self.elapsed_time = elapsed_time
#
# The worker thread class that handles the actual running of a given test.
# Once spawned, it receives requests for work through its input_queue and
# ships the results back through the output_queue.
#
class worker_thread(threading.Thread):
def __init__(self, input_queue, output_queue):
threading.Thread.__init__(self)
self.input_queue = input_queue
self.output_queue = output_queue
def run(self):
while True:
job = self.input_queue.get()
#
# Worker threads continue running until explicitly told to stop with
# a special job.
#
if job.is_break:
return
#
# If the global interrupt handler sets the thread_exit variable,
# we stop doing real work and just report back a "break" in the
# normal command processing has happened.
#
if thread_exit == True:
job.set_is_break(True)
self.output_queue.put(job)
continue
#
# If we are actually supposed to skip this job, do so. Note that
# if is_skip is true, returncode is undefined.
#
if job.is_skip:
if options.verbose:
print("Skip %s" % job.shell_command)
self.output_queue.put(job)
continue
#
# Otherwise go about the business of running tests as normal.
#
else:
if options.verbose:
print("Launch %s" % job.shell_command)
if job.is_example or job.is_pyexample:
#
# If we have an example, the shell command is all we need to
# know. It will be something like "examples/udp/udp-echo" or
# "examples/wireless/mixed-wireless.py"
#
(job.returncode, standard_out, standard_err, et) = run_job_synchronously(job.shell_command,
job.cwd, options.valgrind, job.is_pyexample, job.build_path)
else:
#
# If we're a test suite, we need to provide a little more info
# to the test runner, specifically the base directory and temp
# file name
#
if options.update_data:
update_data = '--update-data'
else:
update_data = ''
(job.returncode, standard_out, standard_err, et) = run_job_synchronously(job.shell_command +
" --xml --tempdir=%s --out=%s %s" % (job.tempdir, job.tmp_file_name, update_data),
job.cwd, options.valgrind, False)
job.set_elapsed_time(et)
if options.verbose:
print("returncode = %d" % job.returncode)
print("---------- begin standard out ----------")
print(standard_out)
print("---------- begin standard err ----------")
print(standard_err)
print("---------- end standard err ----------")
self.output_queue.put(job)
#
# This is the main function that does the work of interacting with the
# test-runner itself.
#
def run_tests():
#
# Pull some interesting configuration information out of waf, primarily
# so we can know where executables can be found, but also to tell us what
# pieces of the system have been built. This will tell us what examples
# are runnable.
#
read_waf_config()
#
# Set the proper suffix.
#
global BUILD_PROFILE_SUFFIX
if BUILD_PROFILE == 'release':
BUILD_PROFILE_SUFFIX = ""
else:
BUILD_PROFILE_SUFFIX = "-" + BUILD_PROFILE
#
# Add the proper prefix and suffix to the test-runner name to
# match what is done in the wscript file.
#
test_runner_name = "%s%s-%s%s" % (APPNAME, VERSION, "test-runner", BUILD_PROFILE_SUFFIX)
#
# Run waf to make sure that everything is built, configured and ready to go
# unless we are explicitly told not to. We want to be careful about causing
# our users pain while waiting for extraneous stuff to compile and link, so
# we allow users that know what they''re doing to not invoke waf at all.
#
if not options.nowaf:
#
# If the user is running the "kinds" or "list" options, there is an
# implied dependency on the test-runner since we call that program
# if those options are selected. We will exit after processing those
# options, so if we see them, we can safely only build the test-runner.
#
# If the user has constrained us to running only a particular type of
# file, we can only ask waf to build what we know will be necessary.
# For example, if the user only wants to run BVT tests, we only have
# to build the test-runner and can ignore all of the examples.
#
# If the user only wants to run a single example, then we can just build
# that example.
#
# If there is no constraint, then we have to build everything since the
# user wants to run everything.
#
if options.kinds or options.list or (len(options.constrain) and options.constrain in core_kinds):
if sys.platform == "win32":
waf_cmd = sys.executable + " waf --target=test-runner"
else:
waf_cmd = sys.executable + " waf --target=test-runner"
elif len(options.example):
if sys.platform == "win32": #Modify for windows
waf_cmd = sys.executable + " waf --target=%s" % os.path.basename(options.example)
else:
waf_cmd = sys.executable + " waf --target=%s" % os.path.basename(options.example)
else:
if sys.platform == "win32": #Modify for windows
waf_cmd = sys.executable + " waf"
else:
waf_cmd = sys.executable + " waf"
if options.verbose:
print("Building: %s" % waf_cmd)
proc = subprocess.Popen(waf_cmd, shell = True)
proc.communicate()
if proc.returncode:
print("Waf died. Not running tests", file=sys.stderr)
return proc.returncode
#
# Dynamically set up paths.
#
make_paths()
#
# Get the information from the build status file.
#
build_status_file = os.path.join (NS3_BUILDDIR, 'build-status.py')
if os.path.exists(build_status_file):
ns3_runnable_programs = get_list_from_file(build_status_file, "ns3_runnable_programs")
ns3_runnable_scripts = get_list_from_file(build_status_file, "ns3_runnable_scripts")
else:
print('The build status file was not found. You must do waf build before running test.py.', file=sys.stderr)
sys.exit(2)
#
# Make a dictionary that maps the name of a program to its path.
#
ns3_runnable_programs_dictionary = {}
for program in ns3_runnable_programs:
# Remove any directory names from path.
program_name = os.path.basename(program)
ns3_runnable_programs_dictionary[program_name] = program
# Generate the lists of examples to run as smoke tests in order to
# ensure that they remain buildable and runnable over time.
#
example_tests = []
example_names_original = []
python_tests = []
for directory in EXAMPLE_DIRECTORIES:
# Set the directories and paths for this example.
example_directory = os.path.join("examples", directory)
examples_to_run_path = os.path.join(example_directory, "examples-to-run.py")
cpp_executable_dir = os.path.join(NS3_BUILDDIR, example_directory)
python_script_dir = os.path.join(example_directory)
# Parse this example directory's file.
parse_examples_to_run_file(
examples_to_run_path,
cpp_executable_dir,
python_script_dir,
example_tests,
example_names_original,
python_tests)
for module in NS3_ENABLED_MODULES:
# Remove the "ns3-" from the module name.
module = module[len("ns3-"):]
# Set the directories and paths for this example.
module_directory = os.path.join("src", module)
example_directory = os.path.join(module_directory, "examples")
examples_to_run_path = os.path.join(module_directory, "test", "examples-to-run.py")
cpp_executable_dir = os.path.join(NS3_BUILDDIR, example_directory)
python_script_dir = os.path.join(example_directory)
# Parse this module's file.
parse_examples_to_run_file(
examples_to_run_path,
cpp_executable_dir,
python_script_dir,
example_tests,
example_names_original,
python_tests)
#
# If lots of logging is enabled, we can crash Python when it tries to
# save all of the text. We just don't allow logging to be turned on when
# test.py runs. If you want to see logging output from your tests, you
# have to run them using the test-runner directly.
#
os.environ["NS_LOG"] = ""
#
# There are a couple of options that imply we can to exit before starting
# up a bunch of threads and running tests. Let's detect these cases and
# handle them without doing all of the hard work.
#
if options.kinds:
path_cmd = os.path.join("utils", test_runner_name + " --print-test-type-list")
(rc, standard_out, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
print(standard_out.decode())
if options.list:
if len(options.constrain):
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list --print-test-types --test-type=%s" % options.constrain)
else:
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list --print-test-types")
(rc, standard_out, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
if rc != 0:
# This is usually a sign that ns-3 crashed or exited uncleanly
print(('test.py error: test-runner return code returned {}'.format(rc)))
print(('To debug, try running {}\n'.format('\'./waf --run \"test-runner --print-test-name-list\"\'')))
return
if isinstance(standard_out, bytes):
standard_out = standard_out.decode()
list_items = standard_out.split('\n')
list_items.sort()
print("Test Type Test Name")
print("--------- ---------")
for item in list_items:
if len(item.strip()):
print(item)
example_names_original.sort()
for item in example_names_original:
print("example ", item)
print()
if options.kinds or options.list:
return
#
# We communicate results in two ways. First, a simple message relating
# PASS, FAIL, CRASH or SKIP is always written to the standard output. It
# is expected that this will be one of the main use cases. A developer can
# just run test.py with no options and see that all of the tests still
# pass.
#
# The second main use case is when detailed status is requested (with the
# --text or --html options). Typicall this will be text if a developer
# finds a problem, or HTML for nightly builds. In these cases, an
# XML file is written containing the status messages from the test suites.
# This file is then read and translated into text or HTML. It is expected
# that nobody will really be interested in the XML, so we write it somewhere
# with a unique name (time) to avoid collisions. In case an error happens, we
# provide a runtime option to retain the temporary files.
#
# When we run examples as smoke tests, they are going to want to create
# lots and lots of trace files. We aren't really interested in the contents
# of the trace files, so we also just stash them off in the temporary dir.
# The retain option also causes these unchecked trace files to be kept.
#
date_and_time = time.strftime("%Y-%m-%d-%H-%M-%S-CUT", time.gmtime())
if not os.path.exists(TMP_OUTPUT_DIR):
os.makedirs(TMP_OUTPUT_DIR)
testpy_output_dir = os.path.join(TMP_OUTPUT_DIR, date_and_time);
if not os.path.exists(testpy_output_dir):
os.makedirs(testpy_output_dir)
#
# Create the main output file and start filling it with XML. We need to
# do this since the tests will just append individual results to this file.
#
xml_results_file = os.path.join(testpy_output_dir, "results.xml")
f = open(xml_results_file, 'w')
f.write('<?xml version="1.0"?>\n')
f.write('<Results>\n')
f.close()
#
# We need to figure out what test suites to execute. We are either given one
# suite or example explicitly via the --suite or --example/--pyexample option,
# or we need to call into the test runner and ask it to list all of the available
# test suites. Further, we need to provide the constraint information if it
# has been given to us.
#
# This translates into allowing the following options with respect to the
# suites
#
# ./test,py: run all of the suites and examples
# ./test.py --constrain=core: run all of the suites of all kinds
# ./test.py --constrain=unit: run all unit suites
# ./test.py --suite=some-test-suite: run a single suite
# ./test.py --example=examples/udp/udp-echo: run single example
# ./test.py --pyexample=examples/wireless/mixed-wireless.py: run python example
# ./test.py --suite=some-suite --example=some-example: run the single suite
#
# We can also use the --constrain option to provide an ordering of test
# execution quite easily.
#
# Flag indicating a specific suite was explicitly requested
single_suite = False
if len(options.suite):
# See if this is a valid test suite.
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list")
(rc, suites, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
if isinstance(suites, bytes):
suites = suites.decode()
if options.suite in suites.split('\n'):
suites = options.suite + "\n"
single_suite = True
else:
print('The test suite was not run because an unknown test suite name was requested.', file=sys.stderr)
sys.exit(2)
elif len(options.example) == 0 and len(options.pyexample) == 0:
if len(options.constrain):
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list --test-type=%s" % options.constrain)
(rc, suites, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
else:
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list")
(rc, suites, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
else:
suites = ""
#
# suite_list will either a single test suite name that the user has
# indicated she wants to run or a list of test suites provided by
# the test-runner possibly according to user provided constraints.
# We go through the trouble of setting up the parallel execution
# even in the case of a single suite to avoid having to process the
# results in two different places.
#
if isinstance(suites, bytes):
suites = suites.decode()
suite_list = suites.split('\n')
#
# Performance tests should only be run when they are requested,
# i.e. they are not run by default in test.py.
# If a specific suite was requested we run it, even if
# it is a performance test.
if not single_suite and options.constrain != 'performance':
# Get a list of all of the performance tests.
path_cmd = os.path.join("utils", test_runner_name + " --print-test-name-list --test-type=%s" % "performance")
(rc, performance_tests, standard_err, et) = run_job_synchronously(path_cmd, os.getcwd(), False, False)
if isinstance(performance_tests, bytes):
performance_tests = performance_tests.decode()
performance_test_list = performance_tests.split('\n')
# Remove any performance tests from the suites list.
for performance_test in performance_test_list:
if performance_test in suite_list:
suite_list.remove(performance_test)
# We now have a possibly large number of test suites to run, so we want to
# run them in parallel. We're going to spin up a number of worker threads
# that will run our test jobs for us.
#
input_queue = queue.Queue(0)
output_queue = queue.Queue(0)
jobs = 0
threads=[]
#
# In Python 2.6 you can just use multiprocessing module, but we don't want
# to introduce that dependency yet; so we jump through a few hoops.
#
processors = 1
if sys.platform != "win32":
if 'SC_NPROCESSORS_ONLN'in os.sysconf_names:
processors = os.sysconf('SC_NPROCESSORS_ONLN')
else:
proc = subprocess.Popen("sysctl -n hw.ncpu", shell = True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout_results, stderr_results = proc.communicate()
stdout_results = stdout_results.decode()
stderr_results = stderr_results.decode()
if len(stderr_results) == 0:
processors = int(stdout_results)
#
# Now, spin up one thread per processor which will eventually mean one test
# per processor running concurrently.
#
for i in range(processors):
thread = worker_thread(input_queue, output_queue)
threads.append(thread)
thread.start()
#
# Keep track of some summary statistics
#
total_tests = 0
skipped_tests = 0
skipped_testnames = []
#
# We now have worker threads spun up, and a list of work to do. So, run
# through the list of test suites and dispatch a job to run each one.
#
# Dispatching will run with unlimited speed and the worker threads will
# execute as fast as possible from the queue.
#
# Note that we actually dispatch tests to be skipped, so all of the
# PASS, FAIL, CRASH and SKIP processing is done in the same place.
#
for test in suite_list:
test = test.strip()
if len(test):
job = Job()
job.set_is_example(False)
job.set_is_pyexample(False)
job.set_display_name(test)
job.set_tmp_file_name(os.path.join(testpy_output_dir, "%s.xml" % test))
job.set_cwd(os.getcwd())
job.set_basedir(os.getcwd())
job.set_tempdir(testpy_output_dir)
if (options.multiple):
multiple = ""
else:
multiple = " --stop-on-failure"
if (len(options.fullness)):
fullness = options.fullness.upper()
fullness = " --fullness=%s" % fullness
else:
fullness = " --fullness=QUICK"
path_cmd = os.path.join("utils", test_runner_name + " --test-name=%s%s%s" % (test, multiple, fullness))
job.set_shell_command(path_cmd)
if options.valgrind and test in core_valgrind_skip_tests:
job.set_is_skip(True)
# Skip tests that will fail if NSC is missing.
if not NSC_ENABLED and test in core_nsc_missing_skip_tests:
job.set_is_skip(True)
if options.verbose:
print("Queue %s" % test)
input_queue.put(job)
jobs = jobs + 1
total_tests = total_tests + 1
#
# We've taken care of the discovered or specified test suites. Now we
# have to deal with examples run as smoke tests. We have a list of all of
# the example programs it makes sense to try and run. Each example will
# have a condition associated with it that must evaluate to true for us
# to try and execute it. This is used to determine if the example has
# a dependency that is not satisfied. For example, if an example depends
# on NSC being configured by waf, that example should have a condition
# that evaluates to true if NSC is enabled. For example,
#
# ("tcp-nsc-zoo", "NSC_ENABLED == True"),
#
# In this case, the example "tcp-nsc-zoo" will only be run if we find the
# waf configuration variable "NSC_ENABLED" to be True.
#
# We don't care at all how the trace files come out, so we just write them
# to a single temporary directory.
#
# XXX As it stands, all of the trace files have unique names, and so file
# collisions can only happen if two instances of an example are running in
# two versions of the test.py process concurrently. We may want to create
# uniquely named temporary traces directories to avoid this problem.
#
# We need to figure out what examples to execute. We are either given one
# suite or example explicitly via the --suite or --example option, or we
# need to walk the list of examples looking for available example
# conditions.
#
# This translates into allowing the following options with respect to the
# suites
#
# ./test.py: run all of the examples
# ./test.py --constrain=unit run no examples
# ./test.py --constrain=example run all of the examples
# ./test.py --suite=some-test-suite: run no examples
# ./test.py --example=some-example: run the single example
# ./test.py --suite=some-suite --example=some-example: run the single example
#
#
if len(options.suite) == 0 and len(options.example) == 0 and len(options.pyexample) == 0:
if len(options.constrain) == 0 or options.constrain == "example":
if ENABLE_EXAMPLES:
for name, test, do_run, do_valgrind_run in example_tests:
# Remove any arguments and directory names from test.
test_name = test.split(' ', 1)[0]
test_name = os.path.basename(test_name)
# Don't try to run this example if it isn't runnable.
if test_name in ns3_runnable_programs_dictionary:
if eval(do_run):
job = Job()
job.set_is_example(True)
job.set_is_pyexample(False)
job.set_display_name(name)
job.set_tmp_file_name("")
job.set_cwd(testpy_output_dir)
job.set_basedir(os.getcwd())
job.set_tempdir(testpy_output_dir)
job.set_shell_command(test)
job.set_build_path(options.buildpath)
if options.valgrind and not eval(do_valgrind_run):
job.set_is_skip (True)
if options.verbose:
print("Queue %s" % test)
input_queue.put(job)
jobs = jobs + 1
total_tests = total_tests + 1
elif len(options.example):
# Add the proper prefix and suffix to the example name to
# match what is done in the wscript file.
example_name = "%s%s-%s%s" % (APPNAME, VERSION, options.example, BUILD_PROFILE_SUFFIX)
# Don't try to run this example if it isn't runnable.
if example_name not in ns3_runnable_programs_dictionary:
print("Example %s is not runnable." % example_name)
else:
#
# If you tell me to run an example, I will try and run the example
# irrespective of any condition.
#
example_path = ns3_runnable_programs_dictionary[example_name]
example_path = os.path.abspath(example_path)
job = Job()
job.set_is_example(True)
job.set_is_pyexample(False)
job.set_display_name(example_path)
job.set_tmp_file_name("")
job.set_cwd(testpy_output_dir)
job.set_basedir(os.getcwd())
job.set_tempdir(testpy_output_dir)
job.set_shell_command(example_path)
job.set_build_path(options.buildpath)
if options.verbose:
print("Queue %s" % example_name)
input_queue.put(job)
jobs = jobs + 1
total_tests = total_tests + 1
#
# Run some Python examples as smoke tests. We have a list of all of
# the example programs it makes sense to try and run. Each example will
# have a condition associated with it that must evaluate to true for us
# to try and execute it. This is used to determine if the example has
# a dependency that is not satisfied.
#
# We don't care at all how the trace files come out, so we just write them
# to a single temporary directory.
#
# We need to figure out what python examples to execute. We are either
# given one pyexample explicitly via the --pyexample option, or we
# need to walk the list of python examples
#
# This translates into allowing the following options with respect to the
# suites
#
# ./test.py --constrain=pyexample run all of the python examples
# ./test.py --pyexample=some-example.py: run the single python example
#
if len(options.suite) == 0 and len(options.example) == 0 and len(options.pyexample) == 0:
if len(options.constrain) == 0 or options.constrain == "pyexample":
if ENABLE_EXAMPLES:
for test, do_run in python_tests:
# Remove any arguments and directory names from test.
test_name = test.split(' ', 1)[0]
test_name = os.path.basename(test_name)
# Don't try to run this example if it isn't runnable.
if test_name in ns3_runnable_scripts:
if eval(do_run):
job = Job()
job.set_is_example(False)
job.set_is_pyexample(True)
job.set_display_name(test)
job.set_tmp_file_name("")
job.set_cwd(testpy_output_dir)
job.set_basedir(os.getcwd())
job.set_tempdir(testpy_output_dir)
job.set_shell_command(test)
job.set_build_path("")
#
# Python programs and valgrind do not work and play
# well together, so we skip them under valgrind.
# We go through the trouble of doing all of this
# work to report the skipped tests in a consistent
# way throught the output formatter.
#
if options.valgrind:
job.set_is_skip (True)
#
# The user can disable python bindings, so we need
# to pay attention to that and give some feedback
# that we're not testing them
#
if not ENABLE_PYTHON_BINDINGS:
job.set_is_skip (True)
if options.verbose:
print("Queue %s" % test)
input_queue.put(job)
jobs = jobs + 1
total_tests = total_tests + 1
elif len(options.pyexample):
# Don't try to run this example if it isn't runnable.
example_name = os.path.basename(options.pyexample)
if example_name not in ns3_runnable_scripts:
print("Example %s is not runnable." % example_name)
else:
#
# If you tell me to run a python example, I will try and run the example
# irrespective of any condition.
#
job = Job()
job.set_is_pyexample(True)
job.set_display_name(options.pyexample)
job.set_tmp_file_name("")
job.set_cwd(testpy_output_dir)
job.set_basedir(os.getcwd())
job.set_tempdir(testpy_output_dir)
job.set_shell_command(options.pyexample)
job.set_build_path("")
if options.verbose:
print("Queue %s" % options.pyexample)
input_queue.put(job)
jobs = jobs + 1
total_tests = total_tests + 1
#
# Tell the worker threads to pack up and go home for the day. Each one
# will exit when they see their is_break task.
#
for i in range(processors):
job = Job()
job.set_is_break(True)
input_queue.put(job)
#
# Now all of the tests have been dispatched, so all we have to do here
# in the main thread is to wait for them to complete. Keyboard interrupt
# handling is broken as mentioned above. We use a signal handler to catch
# sigint and set a global variable. When the worker threads sense this
# they stop doing real work and will just start throwing jobs back at us
# with is_break set to True. In this case, there are no real results so we
# ignore them. If there are real results, we always print PASS or FAIL to
# standard out as a quick indication of what happened.
#
passed_tests = 0
failed_tests = 0
failed_testnames = []
crashed_tests = 0
crashed_testnames = []
valgrind_errors = 0
valgrind_testnames = []
for i in range(jobs):
job = output_queue.get()
if job.is_break:
continue
if job.is_example or job.is_pyexample:
kind = "Example"
else:
kind = "TestSuite"
if job.is_skip:
status = "SKIP"
skipped_tests = skipped_tests + 1
skipped_testnames.append(job.display_name)
else:
if job.returncode == 0:
status = "PASS"
passed_tests = passed_tests + 1
elif job.returncode == 1:
failed_tests = failed_tests + 1
failed_testnames.append(job.display_name)
status = "FAIL"
elif job.returncode == 2:
valgrind_errors = valgrind_errors + 1
valgrind_testnames.append(job.display_name)
status = "VALGR"
else:
crashed_tests = crashed_tests + 1
crashed_testnames.append(job.display_name)
status = "CRASH"
if options.duration or options.constrain == "performance":
print("%s (%.3f): %s %s" % (status, job.elapsed_time, kind, job.display_name))
else:
print("%s: %s %s" % (status, kind, job.display_name))
if job.is_example or job.is_pyexample:
#
# Examples are the odd man out here. They are written without any
# knowledge that they are going to be run as a test, so we need to
# cook up some kind of output for them. We're writing an xml file,
# so we do some simple XML that says we ran the example.
#
# XXX We could add some timing information to the examples, i.e. run
# them through time and print the results here.
#
f = open(xml_results_file, 'a')
f.write('<Example>\n')
example_name = " <Name>%s</Name>\n" % job.display_name
f.write(example_name)
if status == "PASS":
f.write(' <Result>PASS</Result>\n')
elif status == "FAIL":
f.write(' <Result>FAIL</Result>\n')
elif status == "VALGR":
f.write(' <Result>VALGR</Result>\n')
elif status == "SKIP":
f.write(' <Result>SKIP</Result>\n')
else:
f.write(' <Result>CRASH</Result>\n')
f.write(' <Time real="%.3f"/>\n' % job.elapsed_time)
f.write('</Example>\n')
f.close()
else:
#
# If we're not running an example, we're running a test suite.
# These puppies are running concurrently and generating output
# that was written to a temporary file to avoid collisions.
#
# Now that we are executing sequentially in the main thread, we can
# concatenate the contents of the associated temp file to the main
# results file and remove that temp file.
#
# One thing to consider is that a test suite can crash just as
# well as any other program, so we need to deal with that
# possibility as well. If it ran correctly it will return 0
# if it passed, or 1 if it failed. In this case, we can count
# on the results file it saved being complete. If it crashed, it
# will return some other code, and the file should be considered
# corrupt and useless. If the suite didn't create any XML, then
# we're going to have to do it ourselves.
#
# Another issue is how to deal with a valgrind error. If we run
# a test suite under valgrind and it passes, we will get a return
# code of 0 and there will be a valid xml results file since the code
# ran to completion. If we get a return code of 1 under valgrind,
# the test case failed, but valgrind did not find any problems so the
# test case return code was passed through. We will have a valid xml
# results file here as well since the test suite ran. If we see a
# return code of 2, this means that valgrind found an error (we asked
# it to return 2 if it found a problem in run_job_synchronously) but
# the suite ran to completion so there is a valid xml results file.
# If the suite crashes under valgrind we will see some other error
# return code (like 139). If valgrind finds an illegal instruction or
# some other strange problem, it will die with its own strange return
# code (like 132). However, if the test crashes by itself, not under
# valgrind we will also see some other return code.
#
# If the return code is 0, 1, or 2, we have a valid xml file. If we
# get another return code, we have no xml and we can't really say what
# happened -- maybe the TestSuite crashed, maybe valgrind crashed due
# to an illegal instruction. If we get something beside 0-2, we assume
# a crash and fake up an xml entry. After this is all done, we still
# need to indicate a valgrind error somehow, so we fake up an xml entry
# with a VALGR result. Thus, in the case of a working TestSuite that
# fails valgrind, we'll see the PASS entry for the working TestSuite
# followed by a VALGR failing test suite of the same name.
#
if job.is_skip:
f = open(xml_results_file, 'a')
f.write("<Test>\n")
f.write(" <Name>%s</Name>\n" % job.display_name)
f.write(' <Result>SKIP</Result>\n')
f.write("</Test>\n")
f.close()
else:
if job.returncode == 0 or job.returncode == 1 or job.returncode == 2:
f_to = open(xml_results_file, 'a')
f_from = open(job.tmp_file_name)
f_to.write(f_from.read())
f_to.close()
f_from.close()
else:
f = open(xml_results_file, 'a')
f.write("<Test>\n")
f.write(" <Name>%s</Name>\n" % job.display_name)
f.write(' <Result>CRASH</Suite>\n')
f.write("</Test>\n")
f.close()
if job.returncode == 2:
f = open(xml_results_file, 'a')
f.write("<Test>\n")
f.write(" <Name>%s</Name>\n" % job.display_name)
f.write(' <Result>VALGR</Result>\n')
f.write("</Test>\n")
f.close()
#
# We have all of the tests run and the results written out. One final
# bit of housekeeping is to wait for all of the threads to close down
# so we can exit gracefully.
#
for thread in threads:
thread.join()
#
# Back at the beginning of time, we started the body of an XML document
# since the test suites and examples were going to just write their
# individual pieces. So, we need to finish off and close out the XML
# document
#
f = open(xml_results_file, 'a')
f.write('</Results>\n')
f.close()
#
# Print a quick summary of events
#
print("%d of %d tests passed (%d passed, %d skipped, %d failed, %d crashed, %d valgrind errors)" % (passed_tests,
total_tests, passed_tests, skipped_tests, failed_tests, crashed_tests, valgrind_errors))
#
# Repeat summary of skipped, failed, crashed, valgrind events
#
if skipped_testnames:
skipped_testnames.sort()
print('List of SKIPped tests:\n %s' % '\n '.join(map(str, skipped_testnames)))
if failed_testnames:
failed_testnames.sort()
print('List of FAILed tests:\n %s' % '\n '.join(map(str, failed_testnames)))
if crashed_testnames:
crashed_testnames.sort()
print('List of CRASHed tests:\n %s' % '\n '.join(map(str, crashed_testnames)))
if valgrind_testnames:
valgrind_testnames.sort()
print('List of VALGR failures:\n %s' % '\n '.join(map(str, valgrind_testnames)))
#
# The last things to do are to translate the XML results file to "human
# readable form" if the user asked for it (or make an XML file somewhere)
#
if len(options.html):
translate_to_html(xml_results_file, options.html)
if len(options.text):
translate_to_text(xml_results_file, options.text)
if len(options.xml):
shutil.copyfile(xml_results_file, options.xml)
#
# Let the user know if they need to turn on tests or examples.
#
if not ENABLE_TESTS or not ENABLE_EXAMPLES:
print()
if not ENABLE_TESTS:
print('*** Note: ns-3 tests are currently disabled. Enable them by adding')
print('*** "--enable-tests" to ./waf configure or modifying your .ns3rc file.')
print()
if not ENABLE_EXAMPLES:
print('*** Note: ns-3 examples are currently disabled. Enable them by adding')
print('*** "--enable-examples" to ./waf configure or modifying your .ns3rc file.')
print()
#
# Let the user know if they tried to use valgrind but it was not
# present on their machine.
#
if options.valgrind and not VALGRIND_FOUND:
print()
print('*** Note: you are trying to use valgrind, but valgrind could not be found')
print('*** on your machine. All tests and examples will crash or be skipped.')
print()
#
# If we have been asked to retain all of the little temporary files, we
# don't delete tm. If we do delete the temporary files, delete only the
# directory we just created. We don't want to happily delete any retained
# directories, which will probably surprise the user.
#
if not options.retain:
shutil.rmtree(testpy_output_dir)
if passed_tests + skipped_tests == total_tests:
return 0 # success
else:
return 1 # catchall for general errors
def main(argv):
parser = optparse.OptionParser()
parser.add_option("-b", "--buildpath", action="store", type="string", dest="buildpath", default="",
metavar="BUILDPATH",
help="specify the path where ns-3 was built (defaults to the build directory for the current variant)")
parser.add_option("-c", "--constrain", action="store", type="string", dest="constrain", default="",
metavar="KIND",
help="constrain the test-runner by kind of test")
parser.add_option("-d", "--duration", action="store_true", dest="duration", default=False,
help="print the duration of each test suite and example")
parser.add_option("-e", "--example", action="store", type="string", dest="example", default="",
metavar="EXAMPLE",
help="specify a single example to run (no relative path is needed)")
parser.add_option("-u", "--update-data", action="store_true", dest="update_data", default=False,
help="If examples use reference data files, get them to re-generate them")
parser.add_option("-f", "--fullness", action="store", type="string", dest="fullness", default="QUICK",
metavar="FULLNESS",
help="choose the duration of tests to run: QUICK, EXTENSIVE, or TAKES_FOREVER, where EXTENSIVE includes QUICK and TAKES_FOREVER includes QUICK and EXTENSIVE (only QUICK tests are run by default)")
parser.add_option("-g", "--grind", action="store_true", dest="valgrind", default=False,
help="run the test suites and examples using valgrind")
parser.add_option("-k", "--kinds", action="store_true", dest="kinds", default=False,
help="print the kinds of tests available")
parser.add_option("-l", "--list", action="store_true", dest="list", default=False,
help="print the list of known tests")
parser.add_option("-m", "--multiple", action="store_true", dest="multiple", default=False,
help="report multiple failures from test suites and test cases")
parser.add_option("-n", "--nowaf", action="store_true", dest="nowaf", default=False,
help="do not run waf before starting testing")
parser.add_option("-p", "--pyexample", action="store", type="string", dest="pyexample", default="",
metavar="PYEXAMPLE",
help="specify a single python example to run (with relative path)")
parser.add_option("-r", "--retain", action="store_true", dest="retain", default=False,
help="retain all temporary files (which are normally deleted)")
parser.add_option("-s", "--suite", action="store", type="string", dest="suite", default="",
metavar="TEST-SUITE",
help="specify a single test suite to run")
parser.add_option("-t", "--text", action="store", type="string", dest="text", default="",
metavar="TEXT-FILE",
help="write detailed test results into TEXT-FILE.txt")
parser.add_option("-v", "--verbose", action="store_true", dest="verbose", default=False,
help="print progress and informational messages")
parser.add_option("-w", "--web", "--html", action="store", type="string", dest="html", default="",
metavar="HTML-FILE",
help="write detailed test results into HTML-FILE.html")
parser.add_option("-x", "--xml", action="store", type="string", dest="xml", default="",
metavar="XML-FILE",
help="write detailed test results into XML-FILE.xml")
global options
options = parser.parse_args()[0]
signal.signal(signal.SIGINT, sigint_hook)
return run_tests()
if __name__ == '__main__':
sys.exit(main(sys.argv))
|