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
|
// Copyright (c) 2006, Google Inc.
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following disclaimer
// in the documentation and/or other materials provided with the
// distribution.
// * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// Author: Satoru Takabayashi
//
// For reference check out:
// http://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangling
//
// Note that we only have partial C++0x support yet.
#include "demangle.h"
#if defined(GLOG_OS_WINDOWS)
#include <dbghelp.h>
#else
#include <cstdint>
#include <cstdio>
#include <limits>
#endif
_START_GOOGLE_NAMESPACE_
#if !defined(GLOG_OS_WINDOWS)
typedef struct {
const char *abbrev;
const char *real_name;
// Number of arguments in <expression> context, or 0 if disallowed.
int arity;
} AbbrevPair;
// List of operators from Itanium C++ ABI.
static const AbbrevPair kOperatorList[] = {
// New has special syntax (not currently supported).
{"nw", "new", 0},
{"na", "new[]", 0},
// Works except that the 'gs' prefix is not supported.
{"dl", "delete", 1},
{"da", "delete[]", 1},
{"ps", "+", 1}, // "positive"
{"ng", "-", 1}, // "negative"
{"ad", "&", 1}, // "address-of"
{"de", "*", 1}, // "dereference"
{"co", "~", 1},
{"pl", "+", 2},
{"mi", "-", 2},
{"ml", "*", 2},
{"dv", "/", 2},
{"rm", "%", 2},
{"an", "&", 2},
{"or", "|", 2},
{"eo", "^", 2},
{"aS", "=", 2},
{"pL", "+=", 2},
{"mI", "-=", 2},
{"mL", "*=", 2},
{"dV", "/=", 2},
{"rM", "%=", 2},
{"aN", "&=", 2},
{"oR", "|=", 2},
{"eO", "^=", 2},
{"ls", "<<", 2},
{"rs", ">>", 2},
{"lS", "<<=", 2},
{"rS", ">>=", 2},
{"eq", "==", 2},
{"ne", "!=", 2},
{"lt", "<", 2},
{"gt", ">", 2},
{"le", "<=", 2},
{"ge", ">=", 2},
{"nt", "!", 1},
{"aa", "&&", 2},
{"oo", "||", 2},
{"pp", "++", 1},
{"mm", "--", 1},
{"cm", ",", 2},
{"pm", "->*", 2},
{"pt", "->", 0}, // Special syntax
{"cl", "()", 0}, // Special syntax
{"ix", "[]", 2},
{"qu", "?", 3},
{"st", "sizeof", 0}, // Special syntax
{"sz", "sizeof", 1}, // Not a real operator name, but used in expressions.
{nullptr, nullptr, 0},
};
// List of builtin types from Itanium C++ ABI.
//
// Invariant: only one- or two-character type abbreviations here.
static const AbbrevPair kBuiltinTypeList[] = {
{"v", "void", 0},
{"w", "wchar_t", 0},
{"b", "bool", 0},
{"c", "char", 0},
{"a", "signed char", 0},
{"h", "unsigned char", 0},
{"s", "short", 0},
{"t", "unsigned short", 0},
{"i", "int", 0},
{"j", "unsigned int", 0},
{"l", "long", 0},
{"m", "unsigned long", 0},
{"x", "long long", 0},
{"y", "unsigned long long", 0},
{"n", "__int128", 0},
{"o", "unsigned __int128", 0},
{"f", "float", 0},
{"d", "double", 0},
{"e", "long double", 0},
{"g", "__float128", 0},
{"z", "ellipsis", 0},
{"De", "decimal128", 0}, // IEEE 754r decimal floating point (128 bits)
{"Dd", "decimal64", 0}, // IEEE 754r decimal floating point (64 bits)
{"Dc", "decltype(auto)", 0},
{"Da", "auto", 0},
{"Dn", "std::nullptr_t", 0}, // i.e., decltype(nullptr)
{"Df", "decimal32", 0}, // IEEE 754r decimal floating point (32 bits)
{"Di", "char32_t", 0},
{"Du", "char8_t", 0},
{"Ds", "char16_t", 0},
{"Dh", "float16", 0}, // IEEE 754r half-precision float (16 bits)
{nullptr, nullptr, 0},
};
// List of substitutions Itanium C++ ABI.
static const AbbrevPair kSubstitutionList[] = {
{"St", "", 0},
{"Sa", "allocator", 0},
{"Sb", "basic_string", 0},
// std::basic_string<char, std::char_traits<char>,std::allocator<char> >
{"Ss", "string", 0},
// std::basic_istream<char, std::char_traits<char> >
{"Si", "istream", 0},
// std::basic_ostream<char, std::char_traits<char> >
{"So", "ostream", 0},
// std::basic_iostream<char, std::char_traits<char> >
{"Sd", "iostream", 0},
{nullptr, nullptr, 0},
};
// State needed for demangling. This struct is copied in almost every stack
// frame, so every byte counts.
typedef struct {
int mangled_idx; // Cursor of mangled name.
int out_cur_idx; // Cursor of output string.
int prev_name_idx; // For constructors/destructors.
unsigned int prev_name_length : 16; // For constructors/destructors.
signed int nest_level : 15; // For nested names.
unsigned int append : 1; // Append flag.
// Note: for some reason MSVC can't pack "bool append : 1" into the same int
// with the above two fields, so we use an int instead. Amusingly it can pack
// "signed bool" as expected, but relying on that to continue to be a legal
// type seems ill-advised (as it's illegal in at least clang).
} ParseState;
static_assert(sizeof(ParseState) == 4 * sizeof(int),
"unexpected size of ParseState");
// One-off state for demangling that's not subject to backtracking -- either
// constant data, data that's intentionally immune to backtracking (steps), or
// data that would never be changed by backtracking anyway (recursion_depth).
//
// Only one copy of this exists for each call to Demangle, so the size of this
// struct is nearly inconsequential.
typedef struct {
const char* mangled_begin; // Beginning of input string.
char* out; // Beginning of output string.
int out_end_idx; // One past last allowed output character.
int recursion_depth; // For stack exhaustion prevention.
int steps; // Cap how much work we'll do, regardless of depth.
ParseState parse_state; // Backtrackable state copied for most frames.
} State;
namespace {
// Prevent deep recursion / stack exhaustion.
// Also prevent unbounded handling of complex inputs.
class ComplexityGuard {
public:
explicit ComplexityGuard(State* state) : state_(state) {
++state->recursion_depth;
++state->steps;
}
~ComplexityGuard() { --state_->recursion_depth; }
// 256 levels of recursion seems like a reasonable upper limit on depth.
// 128 is not enough to demagle synthetic tests from demangle_unittest.txt:
// "_ZaaZZZZ..." and "_ZaaZcvZcvZ..."
static constexpr int kRecursionDepthLimit = 256;
// We're trying to pick a charitable upper-limit on how many parse steps are
// necessary to handle something that a human could actually make use of.
// This is mostly in place as a bound on how much work we'll do if we are
// asked to demangle an mangled name from an untrusted source, so it should be
// much larger than the largest expected symbol, but much smaller than the
// amount of work we can do in, e.g., a second.
//
// Some real-world symbols from an arbitrary binary started failing between
// 2^12 and 2^13, so we multiply the latter by an extra factor of 16 to set
// the limit.
//
// Spending one second on 2^17 parse steps would require each step to take
// 7.6us, or ~30000 clock cycles, so it's safe to say this can be done in
// under a second.
static constexpr int kParseStepsLimit = 1 << 17;
bool IsTooComplex() const {
return state_->recursion_depth > kRecursionDepthLimit ||
state_->steps > kParseStepsLimit;
}
private:
State* state_;
};
} // namespace
// We don't use strlen() in libc since it's not guaranteed to be async
// signal safe.
static size_t StrLen(const char *str) {
size_t len = 0;
while (*str != '\0') {
++str;
++len;
}
return len;
}
// Returns true if "str" has at least "n" characters remaining.
static bool AtLeastNumCharsRemaining(const char* str, size_t n) {
for (size_t i = 0; i < n; ++i) {
if (str[i] == '\0') {
return false;
}
}
return true;
}
// Returns true if "str" has "prefix" as a prefix.
static bool StrPrefix(const char *str, const char *prefix) {
size_t i = 0;
while (str[i] != '\0' && prefix[i] != '\0' && str[i] == prefix[i]) {
++i;
}
return prefix[i] == '\0'; // Consumed everything in "prefix".
}
static void InitState(State* state,
const char* mangled,
char* out,
size_t out_size) {
state->mangled_begin = mangled;
state->out = out;
state->out_end_idx = static_cast<int>(out_size);
state->recursion_depth = 0;
state->steps = 0;
state->parse_state.mangled_idx = 0;
state->parse_state.out_cur_idx = 0;
state->parse_state.prev_name_idx = 0;
state->parse_state.prev_name_length = 0;
state->parse_state.nest_level = -1;
state->parse_state.append = true;
}
static inline const char* RemainingInput(State* state) {
return &state->mangled_begin[state->parse_state.mangled_idx];
}
// Returns true and advances "mangled_idx" if we find "one_char_token"
// at "mangled_idx" position. It is assumed that "one_char_token" does
// not contain '\0'.
static bool ParseOneCharToken(State *state, const char one_char_token) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (RemainingInput(state)[0] == one_char_token) {
++state->parse_state.mangled_idx;
return true;
}
return false;
}
// Returns true and advances "mangled_cur" if we find "two_char_token"
// at "mangled_cur" position. It is assumed that "two_char_token" does
// not contain '\0'.
static bool ParseTwoCharToken(State *state, const char *two_char_token) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (RemainingInput(state)[0] == two_char_token[0] &&
RemainingInput(state)[1] == two_char_token[1]) {
state->parse_state.mangled_idx += 2;
return true;
}
return false;
}
// Returns true and advances "mangled_cur" if we find any character in
// "char_class" at "mangled_cur" position.
static bool ParseCharClass(State *state, const char *char_class) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (RemainingInput(state)[0] == '\0') {
return false;
}
const char *p = char_class;
for (; *p != '\0'; ++p) {
if (RemainingInput(state)[0] == *p) {
++state->parse_state.mangled_idx;
return true;
}
}
return false;
}
static bool ParseDigit(State* state, int* digit) {
char c = RemainingInput(state)[0];
if (ParseCharClass(state, "0123456789")) {
if (digit != nullptr) {
*digit = c - '0';
}
return true;
}
return false;
}
// This function is used for handling an optional non-terminal.
static bool Optional(bool /*status*/) {
return true;
}
// This function is used for handling <non-terminal>+ syntax.
typedef bool (*ParseFunc)(State *);
static bool OneOrMore(ParseFunc parse_func, State *state) {
if (parse_func(state)) {
while (parse_func(state)) {
}
return true;
}
return false;
}
// This function is used for handling <non-terminal>* syntax. The function
// always returns true and must be followed by a termination token or a
// terminating sequence not handled by parse_func (e.g.
// ParseOneCharToken(state, 'E')).
static bool ZeroOrMore(ParseFunc parse_func, State *state) {
while (parse_func(state)) {
}
return true;
}
// Append "str" at "out_cur_idx". If there is an overflow, out_cur_idx is
// set to out_end_idx+1. The output string is ensured to
// always terminate with '\0' as long as there is no overflow.
static void Append(State* state, const char* const str, const size_t length) {
for (size_t i = 0; i < length; ++i) {
if (state->parse_state.out_cur_idx + 1 <
state->out_end_idx) { // +1 for '\0'
state->out[state->parse_state.out_cur_idx++] = str[i];
} else {
// signal overflow
state->parse_state.out_cur_idx = state->out_end_idx + 1;
break;
}
}
if (state->parse_state.out_cur_idx < state->out_end_idx) {
state->out[state->parse_state.out_cur_idx] =
'\0'; // Terminate it with '\0'
}
}
// We don't use equivalents in libc to avoid locale issues.
static bool IsLower(char c) {
return c >= 'a' && c <= 'z';
}
static bool IsAlpha(char c) {
return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z');
}
static bool IsDigit(char c) {
return c >= '0' && c <= '9';
}
// Returns true if "str" is a function clone suffix. These suffixes are used
// by GCC 4.5.x and later versions (and our locally-modified version of GCC
// 4.4.x) to indicate functions which have been cloned during optimization.
// We treat any sequence (.<alpha>+.<digit>+)+ as a function clone suffix.
// Additionally, '_' is allowed along with the alphanumeric sequence.
static bool IsFunctionCloneSuffix(const char *str) {
size_t i = 0;
while (str[i] != '\0') {
bool parsed = false;
// Consume a single [.<alpha> | _]*[.<digit>]* sequence.
if (str[i] == '.' && (IsAlpha(str[i + 1]) || str[i + 1] == '_')) {
parsed = true;
i += 2;
while (IsAlpha(str[i]) || str[i] == '_') {
++i;
}
}
if (str[i] == '.' && IsDigit(str[i + 1])) {
parsed = true;
i += 2;
while (IsDigit(str[i])) {
++i;
}
}
if (!parsed) {
return false;
}
}
return true; // Consumed everything in "str".
}
static bool EndsWith(State* state, const char chr) {
return state->parse_state.out_cur_idx > 0 &&
state->parse_state.out_cur_idx < state->out_end_idx &&
chr == state->out[state->parse_state.out_cur_idx - 1];
}
// Append "str" with some tweaks, iff "append" state is true.
static void MaybeAppendWithLength(State* state,
const char* const str,
const size_t length) {
if (state->parse_state.append && length > 0) {
// Append a space if the output buffer ends with '<' and "str"
// starts with '<' to avoid <<<.
if (str[0] == '<' && EndsWith(state, '<')) {
Append(state, " ", 1);
}
// Remember the last identifier name for ctors/dtors,
// but only if we haven't yet overflown the buffer.
if (state->parse_state.out_cur_idx < state->out_end_idx &&
(IsAlpha(str[0]) || str[0] == '_')) {
state->parse_state.prev_name_idx = state->parse_state.out_cur_idx;
state->parse_state.prev_name_length = static_cast<unsigned int>(length);
}
Append(state, str, length);
}
}
// Appends a positive decimal number to the output if appending is enabled.
static bool MaybeAppendDecimal(State* state, int val) {
// Max {32-64}-bit unsigned int is 20 digits.
constexpr size_t kMaxLength = 20;
char buf[kMaxLength];
// We can't use itoa or sprintf as neither is specified to be
// async-signal-safe.
if (state->parse_state.append) {
// We can't have a one-before-the-beginning pointer, so instead start with
// one-past-the-end and manipulate one character before the pointer.
char* p = &buf[kMaxLength];
do { // val=0 is the only input that should write a leading zero digit.
*--p = static_cast<char>((val % 10) + '0');
val /= 10;
} while (p > buf && val != 0);
// 'p' landed on the last character we set. How convenient.
Append(state, p, kMaxLength - static_cast<size_t>(p - buf));
}
return true;
}
// A convenient wrapper around MaybeAppendWithLength().
// Returns true so that it can be placed in "if" conditions.
static bool MaybeAppend(State* state, const char* const str) {
if (state->parse_state.append) {
size_t length = StrLen(str);
MaybeAppendWithLength(state, str, length);
}
return true;
}
// This function is used for handling nested names.
static bool EnterNestedName(State *state) {
state->parse_state.nest_level = 0;
return true;
}
// This function is used for handling nested names.
static bool LeaveNestedName(State* state, int16_t prev_value) {
state->parse_state.nest_level = prev_value;
return true;
}
// Disable the append mode not to print function parameters, etc.
static bool DisableAppend(State *state) {
state->parse_state.append = false;
return true;
}
// Restore the append mode to the previous state.
static bool RestoreAppend(State *state, bool prev_value) {
state->parse_state.append = prev_value;
return true;
}
// Increase the nest level for nested names.
static void MaybeIncreaseNestLevel(State *state) {
if (state->parse_state.nest_level > -1) {
++state->parse_state.nest_level;
}
}
// Appends :: for nested names if necessary.
static void MaybeAppendSeparator(State *state) {
if (state->parse_state.nest_level >= 1) {
MaybeAppend(state, "::");
}
}
// Cancel the last separator if necessary.
static void MaybeCancelLastSeparator(State *state) {
if (state->parse_state.nest_level >= 1 && state->parse_state.append &&
state->parse_state.out_cur_idx >= 2) {
state->parse_state.out_cur_idx -= 2;
state->out[state->parse_state.out_cur_idx] = '\0';
}
}
// Returns true if the identifier of the given length pointed to by
// "mangled_cur" is anonymous namespace.
static bool IdentifierIsAnonymousNamespace(State* state, size_t length) {
// Returns true if "anon_prefix" is a proper prefix of "mangled_cur".
static const char anon_prefix[] = "_GLOBAL__N_";
return (length > (sizeof(anon_prefix) - 1) &&
StrPrefix(RemainingInput(state), anon_prefix));
}
// Forward declarations of our parsing functions.
static bool ParseMangledName(State *state);
static bool ParseEncoding(State *state);
static bool ParseName(State *state);
static bool ParseUnscopedName(State* state);
static bool ParseNestedName(State *state);
static bool ParsePrefix(State *state);
static bool ParseUnqualifiedName(State *state);
static bool ParseSourceName(State *state);
static bool ParseLocalSourceName(State *state);
static bool ParseUnnamedTypeName(State* state);
static bool ParseNumber(State *state, int *number_out);
static bool ParseFloatNumber(State *state);
static bool ParseSeqId(State *state);
static bool ParseIdentifier(State* state, size_t length);
static bool ParseOperatorName(State* state, int* arity);
static bool ParseSpecialName(State *state);
static bool ParseCallOffset(State *state);
static bool ParseNVOffset(State *state);
static bool ParseVOffset(State *state);
static bool ParseAbiTags(State* state);
static bool ParseCtorDtorName(State *state);
static bool ParseDecltype(State* state);
static bool ParseType(State *state);
static bool ParseCVQualifiers(State *state);
static bool ParseBuiltinType(State *state);
static bool ParseFunctionType(State *state);
static bool ParseBareFunctionType(State *state);
static bool ParseClassEnumType(State *state);
static bool ParseArrayType(State *state);
static bool ParsePointerToMemberType(State *state);
static bool ParseTemplateParam(State *state);
static bool ParseTemplateTemplateParam(State *state);
static bool ParseTemplateArgs(State *state);
static bool ParseTemplateArg(State *state);
static bool ParseBaseUnresolvedName(State* state);
static bool ParseUnresolvedName(State* state);
static bool ParseExpression(State *state);
static bool ParseExprPrimary(State *state);
static bool ParseExprCastValue(State* state);
static bool ParseLocalName(State *state);
static bool ParseLocalNameSuffix(State* state);
static bool ParseDiscriminator(State *state);
static bool ParseSubstitution(State* state, bool accept_std);
// Implementation note: the following code is a straightforward
// translation of the Itanium C++ ABI defined in BNF with a couple of
// exceptions.
//
// - Support GNU extensions not defined in the Itanium C++ ABI
// - <prefix> and <template-prefix> are combined to avoid infinite loop
// - Reorder patterns to shorten the code
// - Reorder patterns to give greedier functions precedence
// We'll mark "Less greedy than" for these cases in the code
//
// Each parsing function changes the parse state and returns true on
// success, or returns false and doesn't change the parse state (note:
// the parse-steps counter increases regardless of success or failure).
// To ensure that the parse state isn't changed in the latter case, we
// save the original state before we call multiple parsing functions
// consecutively with &&, and restore it if unsuccessful. See
// ParseEncoding() as an example of this convention. We follow the
// convention throughout the code.
//
// Originally we tried to do demangling without following the full ABI
// syntax but it turned out we needed to follow the full syntax to
// parse complicated cases like nested template arguments. Note that
// implementing a full-fledged demangler isn't trivial (libiberty's
// cp-demangle.c has +4300 lines).
//
// Note that (foo) in <(foo) ...> is a modifier to be ignored.
//
// Reference:
// - Itanium C++ ABI
// <https://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangling>
// <mangled-name> ::= _Z <encoding>
static bool ParseMangledName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
return ParseTwoCharToken(state, "_Z") && ParseEncoding(state);
}
// <encoding> ::= <(function) name> <bare-function-type>
// ::= <(data) name>
// ::= <special-name>
static bool ParseEncoding(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
// Implementing the first two productions together as <name>
// [<bare-function-type>] avoids exponential blowup of backtracking.
//
// Since Optional(...) can't fail, there's no need to copy the state for
// backtracking.
if (ParseName(state) && Optional(ParseBareFunctionType(state))) {
return true;
}
if (ParseSpecialName(state)) {
return true;
}
return false;
}
// <name> ::= <nested-name>
// ::= <unscoped-template-name> <template-args>
// ::= <unscoped-name>
// ::= <local-name>
static bool ParseName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseNestedName(state) || ParseLocalName(state)) {
return true;
}
// We reorganize the productions to avoid re-parsing unscoped names.
// - Inline <unscoped-template-name> productions:
// <name> ::= <substitution> <template-args>
// ::= <unscoped-name> <template-args>
// ::= <unscoped-name>
// - Merge the two productions that start with unscoped-name:
// <name> ::= <unscoped-name> [<template-args>]
ParseState copy = state->parse_state;
// "std<...>" isn't a valid name.
if (ParseSubstitution(state, /*accept_std=*/false) &&
ParseTemplateArgs(state)) {
return true;
}
state->parse_state = copy;
// Note there's no need to restore state after this since only the first
// subparser can fail.
return ParseUnscopedName(state) && Optional(ParseTemplateArgs(state));
}
// <unscoped-name> ::= <unqualified-name>
// ::= St <unqualified-name>
static bool ParseUnscopedName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseUnqualifiedName(state)) {
return true;
}
ParseState copy = state->parse_state;
if (ParseTwoCharToken(state, "St") && MaybeAppend(state, "std::") &&
ParseUnqualifiedName(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <ref-qualifer> ::= R // lvalue method reference qualifier
// ::= O // rvalue method reference qualifier
static inline bool ParseRefQualifier(State* state) {
return ParseCharClass(state, "OR");
}
// <nested-name> ::= N [<CV-qualifiers>] [<ref-qualifier>] <prefix>
// <unqualified-name> E
// ::= N [<CV-qualifiers>] [<ref-qualifier>] <template-prefix>
// <template-args> E
static bool ParseNestedName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'N') && EnterNestedName(state) &&
Optional(ParseCVQualifiers(state)) &&
Optional(ParseRefQualifier(state)) && ParsePrefix(state) &&
LeaveNestedName(state, copy.nest_level) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// This part is tricky. If we literally translate them to code, we'll
// end up infinite loop. Hence we merge them to avoid the case.
//
// <prefix> ::= <prefix> <unqualified-name>
// ::= <template-prefix> <template-args>
// ::= <template-param>
// ::= <substitution>
// ::= # empty
// <template-prefix> ::= <prefix> <(template) unqualified-name>
// ::= <template-param>
// ::= <substitution>
static bool ParsePrefix(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
bool has_something = false;
while (true) {
MaybeAppendSeparator(state);
if (ParseTemplateParam(state) ||
ParseSubstitution(state, /*accept_std=*/true) ||
ParseUnscopedName(state) ||
(ParseOneCharToken(state, 'M') && ParseUnnamedTypeName(state))) {
has_something = true;
MaybeIncreaseNestLevel(state);
continue;
}
MaybeCancelLastSeparator(state);
if (has_something && ParseTemplateArgs(state)) {
return ParsePrefix(state);
} else {
break;
}
}
return true;
}
// <unqualified-name> ::= <operator-name> [<abi-tags>]
// ::= <ctor-dtor-name> [<abi-tags>]
// ::= <source-name> [<abi-tags>]
// ::= <local-source-name> [<abi-tags>]
// ::= <unnamed-type-name> [<abi-tags>]
//
// <local-source-name> is a GCC extension; see below.
static bool ParseUnqualifiedName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseOperatorName(state, nullptr) || ParseCtorDtorName(state) ||
ParseSourceName(state) || ParseLocalSourceName(state) ||
ParseUnnamedTypeName(state)) {
return ParseAbiTags(state);
}
return false;
}
// <abi-tags> ::= <abi-tag> [<abi-tags>]
// <abi-tag> ::= B <source-name>
static bool ParseAbiTags(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
while (ParseOneCharToken(state, 'B')) {
ParseState copy = state->parse_state;
MaybeAppend(state, "[abi:");
if (!ParseSourceName(state)) {
state->parse_state = copy;
return false;
}
MaybeAppend(state, "]");
}
return true;
}
// <source-name> ::= <positive length number> <identifier>
static bool ParseSourceName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
int length = -1;
if (ParseNumber(state, &length) &&
ParseIdentifier(state, static_cast<size_t>(length))) {
return true;
}
state->parse_state = copy;
return false;
}
// <local-source-name> ::= L <source-name> [<discriminator>]
//
// References:
// https://gcc.gnu.org/bugzilla/show_bug.cgi?id=31775
// https://gcc.gnu.org/viewcvs?view=rev&revision=124467
static bool ParseLocalSourceName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'L') && ParseSourceName(state) &&
Optional(ParseDiscriminator(state))) {
return true;
}
state->parse_state = copy;
return false;
}
// <unnamed-type-name> ::= Ut [<(nonnegative) number>] _
// ::= <closure-type-name>
// <closure-type-name> ::= Ul <lambda-sig> E [<(nonnegative) number>] _
// <lambda-sig> ::= <(parameter) type>+
static bool ParseUnnamedTypeName(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
// Type's 1-based index n is encoded as { "", n == 1; itoa(n-2), otherwise }.
// Optionally parse the encoded value into 'which' and add 2 to get the index.
int which = -1;
// Unnamed type local to function or class.
if (ParseTwoCharToken(state, "Ut") && Optional(ParseNumber(state, &which)) &&
which <= std::numeric_limits<int>::max() - 2 && // Don't overflow.
ParseOneCharToken(state, '_')) {
MaybeAppend(state, "{unnamed type#");
MaybeAppendDecimal(state, 2 + which);
MaybeAppend(state, "}");
return true;
}
state->parse_state = copy;
// Closure type.
which = -1;
if (ParseTwoCharToken(state, "Ul") && DisableAppend(state) &&
OneOrMore(ParseType, state) && RestoreAppend(state, copy.append) &&
ParseOneCharToken(state, 'E') && Optional(ParseNumber(state, &which)) &&
which <= std::numeric_limits<int>::max() - 2 && // Don't overflow.
ParseOneCharToken(state, '_')) {
MaybeAppend(state, "{lambda()#");
MaybeAppendDecimal(state, 2 + which);
MaybeAppend(state, "}");
return true;
}
state->parse_state = copy;
return false;
}
// <number> ::= [n] <non-negative decimal integer>
// If "number_out" is non-null, then *number_out is set to the value of the
// parsed number on success.
static bool ParseNumber(State *state, int *number_out) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
bool negative = false;
if (ParseOneCharToken(state, 'n')) {
negative = true;
}
const char* p = RemainingInput(state);
uint64_t number = 0;
for (; *p != '\0'; ++p) {
if (IsDigit(*p)) {
number = number * 10 + static_cast<uint64_t>(*p - '0');
} else {
break;
}
}
// Apply the sign with uint64_t arithmetic so overflows aren't UB. Gives
// "incorrect" results for out-of-range inputs, but negative values only
// appear for literals, which aren't printed.
if (negative) {
number = ~number + 1;
}
if (p != RemainingInput(state)) { // Conversion succeeded.
state->parse_state.mangled_idx += p - RemainingInput(state);
if (number_out != nullptr) {
// Note: possibly truncate "number".
*number_out = static_cast<int>(number);
}
return true;
}
return false;
}
// Floating-point literals are encoded using a fixed-length lowercase
// hexadecimal string.
static bool ParseFloatNumber(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
const char* p = RemainingInput(state);
for (; *p != '\0'; ++p) {
if (!IsDigit(*p) && !(*p >= 'a' && *p <= 'f')) {
break;
}
}
if (p != RemainingInput(state)) { // Conversion succeeded.
state->parse_state.mangled_idx += p - RemainingInput(state);
return true;
}
return false;
}
// The <seq-id> is a sequence number in base 36,
// using digits and upper case letters
static bool ParseSeqId(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
const char* p = RemainingInput(state);
for (; *p != '\0'; ++p) {
if (!IsDigit(*p) && !(*p >= 'A' && *p <= 'Z')) {
break;
}
}
if (p != RemainingInput(state)) { // Conversion succeeded.
state->parse_state.mangled_idx += p - RemainingInput(state);
return true;
}
return false;
}
// <identifier> ::= <unqualified source code identifier> (of given length)
static bool ParseIdentifier(State* state, size_t length) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (!AtLeastNumCharsRemaining(RemainingInput(state), length)) {
return false;
}
if (IdentifierIsAnonymousNamespace(state, length)) {
MaybeAppend(state, "(anonymous namespace)");
} else {
MaybeAppendWithLength(state, RemainingInput(state), length);
}
state->parse_state.mangled_idx += length;
return true;
}
// <operator-name> ::= nw, and other two letters cases
// ::= cv <type> # (cast)
// ::= v <digit> <source-name> # vendor extended operator
static bool ParseOperatorName(State* state, int* arity) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (!AtLeastNumCharsRemaining(RemainingInput(state), 2)) {
return false;
}
// First check with "cv" (cast) case.
ParseState copy = state->parse_state;
if (ParseTwoCharToken(state, "cv") && MaybeAppend(state, "operator ") &&
EnterNestedName(state) && ParseType(state) &&
LeaveNestedName(state, copy.nest_level)) {
if (arity != nullptr) {
*arity = 1;
}
return true;
}
state->parse_state = copy;
// Then vendor extended operators.
if (ParseOneCharToken(state, 'v') && ParseDigit(state, arity) &&
ParseSourceName(state)) {
return true;
}
state->parse_state = copy;
// Other operator names should start with a lower alphabet followed
// by a lower/upper alphabet.
if (!(IsLower(RemainingInput(state)[0]) &&
IsAlpha(RemainingInput(state)[1]))) {
return false;
}
// We may want to perform a binary search if we really need speed.
const AbbrevPair *p;
for (p = kOperatorList; p->abbrev != nullptr; ++p) {
if (RemainingInput(state)[0] == p->abbrev[0] &&
RemainingInput(state)[1] == p->abbrev[1]) {
if (arity != nullptr) {
*arity = p->arity;
}
MaybeAppend(state, "operator");
if (IsLower(*p->real_name)) { // new, delete, etc.
MaybeAppend(state, " ");
}
MaybeAppend(state, p->real_name);
state->parse_state.mangled_idx += 2;
return true;
}
}
return false;
}
// <special-name> ::= TV <type>
// ::= TT <type>
// ::= TI <type>
// ::= TS <type>
// ::= TH <type> # thread-local
// ::= Tc <call-offset> <call-offset> <(base) encoding>
// ::= GV <(object) name>
// ::= T <call-offset> <(base) encoding>
// G++ extensions:
// ::= TC <type> <(offset) number> _ <(base) type>
// ::= TF <type>
// ::= TJ <type>
// ::= GR <name>
// ::= GA <encoding>
// ::= Th <call-offset> <(base) encoding>
// ::= Tv <call-offset> <(base) encoding>
//
// Note: we don't care much about them since they don't appear in
// stack traces. The are special data.
static bool ParseSpecialName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'T') && ParseCharClass(state, "VTISH") &&
ParseType(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "Tc") && ParseCallOffset(state) &&
ParseCallOffset(state) && ParseEncoding(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "GV") && ParseName(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'T') && ParseCallOffset(state) &&
ParseEncoding(state)) {
return true;
}
state->parse_state = copy;
// G++ extensions
if (ParseTwoCharToken(state, "TC") && ParseType(state) &&
ParseNumber(state, nullptr) && ParseOneCharToken(state, '_') &&
DisableAppend(state) && ParseType(state)) {
RestoreAppend(state, copy.append);
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'T') && ParseCharClass(state, "FJ") &&
ParseType(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "GR") && ParseName(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "GA") && ParseEncoding(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'T') && ParseCharClass(state, "hv") &&
ParseCallOffset(state) && ParseEncoding(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <call-offset> ::= h <nv-offset> _
// ::= v <v-offset> _
static bool ParseCallOffset(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'h') && ParseNVOffset(state) &&
ParseOneCharToken(state, '_')) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'v') && ParseVOffset(state) &&
ParseOneCharToken(state, '_')) {
return true;
}
state->parse_state = copy;
return false;
}
// <nv-offset> ::= <(offset) number>
static bool ParseNVOffset(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
return ParseNumber(state, nullptr);
}
// <v-offset> ::= <(offset) number> _ <(virtual offset) number>
static bool ParseVOffset(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseNumber(state, nullptr) && ParseOneCharToken(state, '_') &&
ParseNumber(state, nullptr)) {
return true;
}
state->parse_state = copy;
return false;
}
// <ctor-dtor-name> ::= C1 | C2 | C3 | CI1 <base-class-type> | CI2
// <base-class-type>
// ::= D0 | D1 | D2
// # GCC extensions: "unified" constructor/destructor. See
// #
// https://github.com/gcc-mirror/gcc/blob/7ad17b583c3643bd4557f29b8391ca7ef08391f5/gcc/cp/mangle.c#L1847
// ::= C4 | D4
static bool ParseCtorDtorName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'C')) {
if (ParseCharClass(state, "1234")) {
const char* const prev_name =
state->out + state->parse_state.prev_name_idx;
MaybeAppendWithLength(state, prev_name,
state->parse_state.prev_name_length);
return true;
} else if (ParseOneCharToken(state, 'I') && ParseCharClass(state, "12") &&
ParseClassEnumType(state)) {
return true;
}
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'D') && ParseCharClass(state, "0124")) {
const char* const prev_name = state->out + state->parse_state.prev_name_idx;
MaybeAppend(state, "~");
MaybeAppendWithLength(state, prev_name,
state->parse_state.prev_name_length);
return true;
}
state->parse_state = copy;
return false;
}
// <decltype> ::= Dt <expression> E # decltype of an id-expression or class
// # member access (C++0x)
// ::= DT <expression> E # decltype of an expression (C++0x)
static bool ParseDecltype(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'D') && ParseCharClass(state, "tT") &&
ParseExpression(state) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <type> ::= <CV-qualifiers> <type>
// ::= P <type> # pointer-to
// ::= R <type> # reference-to
// ::= O <type> # rvalue reference-to (C++0x)
// ::= C <type> # complex pair (C 2000)
// ::= G <type> # imaginary (C 2000)
// ::= U <source-name> <type> # vendor extended type qualifier
// ::= <builtin-type>
// ::= <function-type>
// ::= <class-enum-type> # note: just an alias for <name>
// ::= <array-type>
// ::= <pointer-to-member-type>
// ::= <template-template-param> <template-args>
// ::= <template-param>
// ::= <decltype>
// ::= <substitution>
// ::= Dp <type> # pack expansion of (C++0x)
// ::= Dv <num-elems> _ # GNU vector extension
//
static bool ParseType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
// We should check CV-qualifers, and PRGC things first.
//
// CV-qualifiers overlap with some operator names, but an operator name is not
// valid as a type. To avoid an ambiguity that can lead to exponential time
// complexity, refuse to backtrack the CV-qualifiers.
//
// _Z4aoeuIrMvvE
// => _Z 4aoeuI rM v v E
// aoeu<operator%=, void, void>
// => _Z 4aoeuI r Mv v E
// aoeu<void void::* restrict>
//
// By consuming the CV-qualifiers first, the former parse is disabled.
if (ParseCVQualifiers(state)) {
const bool result = ParseType(state);
if (!result) {
state->parse_state = copy;
}
return result;
}
state->parse_state = copy;
// Similarly, these tag characters can overlap with other <name>s resulting in
// two different parse prefixes that land on <template-args> in the same
// place, such as "C3r1xI...". So, disable the "ctor-name = C3" parse by
// refusing to backtrack the tag characters.
if (ParseCharClass(state, "OPRCG")) {
const bool result = ParseType(state);
if (!result) {
state->parse_state = copy;
}
return result;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "Dp") && ParseType(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'U') && ParseSourceName(state) &&
ParseType(state)) {
return true;
}
state->parse_state = copy;
if (ParseBuiltinType(state) || ParseFunctionType(state) ||
ParseClassEnumType(state) || ParseArrayType(state) ||
ParsePointerToMemberType(state) || ParseDecltype(state) ||
// "std" on its own isn't a type.
ParseSubstitution(state, /*accept_std=*/false)) {
return true;
}
if (ParseTemplateTemplateParam(state) && ParseTemplateArgs(state)) {
return true;
}
state->parse_state = copy;
// Less greedy than <template-template-param> <template-args>.
if (ParseTemplateParam(state)) {
return true;
}
if (ParseTwoCharToken(state, "Dv") && ParseNumber(state, nullptr) &&
ParseOneCharToken(state, '_')) {
return true;
}
state->parse_state = copy;
return false;
}
// <CV-qualifiers> ::= [r] [V] [K]
// We don't allow empty <CV-qualifiers> to avoid infinite loop in
// ParseType().
static bool ParseCVQualifiers(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
int num_cv_qualifiers = 0;
num_cv_qualifiers += ParseOneCharToken(state, 'r');
num_cv_qualifiers += ParseOneCharToken(state, 'V');
num_cv_qualifiers += ParseOneCharToken(state, 'K');
return num_cv_qualifiers > 0;
}
// <builtin-type> ::= v, etc. # single-character builtin types
// ::= u <source-name>
// ::= Dd, etc. # two-character builtin types
//
// Not supported:
// ::= DF <number> _ # _FloatN (N bits)
//
static bool ParseBuiltinType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
const AbbrevPair *p;
for (p = kBuiltinTypeList; p->abbrev != nullptr; ++p) {
// Guaranteed only 1- or 2-character strings in kBuiltinTypeList.
if (p->abbrev[1] == '\0') {
if (ParseOneCharToken(state, p->abbrev[0])) {
MaybeAppend(state, p->real_name);
return true;
}
} else if (p->abbrev[2] == '\0' && ParseTwoCharToken(state, p->abbrev)) {
MaybeAppend(state, p->real_name);
return true;
}
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'u') && ParseSourceName(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <exception-spec> ::= Do # non-throwing
// exception-specification (e.g.,
// noexcept, throw())
// ::= DO <expression> E # computed (instantiation-dependent)
// noexcept
// ::= Dw <type>+ E # dynamic exception specification
// with instantiation-dependent types
static bool ParseExceptionSpec(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseTwoCharToken(state, "Do")) {
return true;
}
ParseState copy = state->parse_state;
if (ParseTwoCharToken(state, "DO") && ParseExpression(state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "Dw") && OneOrMore(ParseType, state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <function-type> ::= [exception-spec] F [Y] <bare-function-type> [O] E
static bool ParseFunctionType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (Optional(ParseExceptionSpec(state)) && ParseOneCharToken(state, 'F') &&
Optional(ParseOneCharToken(state, 'Y')) && ParseBareFunctionType(state) &&
Optional(ParseOneCharToken(state, 'O')) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <bare-function-type> ::= <(signature) type>+
static bool ParseBareFunctionType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
DisableAppend(state);
if (OneOrMore(ParseType, state)) {
RestoreAppend(state, copy.append);
MaybeAppend(state, "()");
return true;
}
state->parse_state = copy;
return false;
}
// <class-enum-type> ::= <name>
static bool ParseClassEnumType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
return ParseName(state);
}
// <array-type> ::= A <(positive dimension) number> _ <(element) type>
// ::= A [<(dimension) expression>] _ <(element) type>
static bool ParseArrayType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'A') && ParseNumber(state, nullptr) &&
ParseOneCharToken(state, '_') && ParseType(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'A') && Optional(ParseExpression(state)) &&
ParseOneCharToken(state, '_') && ParseType(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <pointer-to-member-type> ::= M <(class) type> <(member) type>
static bool ParsePointerToMemberType(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'M') && ParseType(state) && ParseType(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <template-param> ::= T_
// ::= T <parameter-2 non-negative number> _
static bool ParseTemplateParam(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseTwoCharToken(state, "T_")) {
MaybeAppend(state, "?"); // We don't support template substitutions.
return true;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'T') && ParseNumber(state, nullptr) &&
ParseOneCharToken(state, '_')) {
MaybeAppend(state, "?"); // We don't support template substitutions.
return true;
}
state->parse_state = copy;
return false;
}
// <template-template-param> ::= <template-param>
// ::= <substitution>
static bool ParseTemplateTemplateParam(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
return (ParseTemplateParam(state) ||
// "std" on its own isn't a template.
ParseSubstitution(state, /*accept_std=*/false));
}
// <template-args> ::= I <template-arg>+ E
static bool ParseTemplateArgs(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
DisableAppend(state);
if (ParseOneCharToken(state, 'I') && OneOrMore(ParseTemplateArg, state) &&
ParseOneCharToken(state, 'E')) {
RestoreAppend(state, copy.append);
MaybeAppend(state, "<>");
return true;
}
state->parse_state = copy;
return false;
}
// <template-arg> ::= <type>
// ::= <expr-primary>
// ::= J <template-arg>* E # argument pack
// ::= X <expression> E
static bool ParseTemplateArg(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'J') && ZeroOrMore(ParseTemplateArg, state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
// There can be significant overlap between the following leading to
// exponential backtracking:
//
// <expr-primary> ::= L <type> <expr-cast-value> E
// e.g. L 2xxIvE 1 E
// <type> ==> <local-source-name> <template-args>
// e.g. L 2xx IvE
//
// This means parsing an entire <type> twice, and <type> can contain
// <template-arg>, so this can generate exponential backtracking. There is
// only overlap when the remaining input starts with "L <source-name>", so
// parse all cases that can start this way jointly to share the common prefix.
//
// We have:
//
// <template-arg> ::= <type>
// ::= <expr-primary>
//
// First, drop all the productions of <type> that must start with something
// other than 'L'. All that's left is <class-enum-type>; inline it.
//
// <type> ::= <nested-name> # starts with 'N'
// ::= <unscoped-name>
// ::= <unscoped-template-name> <template-args>
// ::= <local-name> # starts with 'Z'
//
// Drop and inline again:
//
// <type> ::= <unscoped-name>
// ::= <unscoped-name> <template-args>
// ::= <substitution> <template-args> # starts with 'S'
//
// Merge the first two, inline <unscoped-name>, drop last:
//
// <type> ::= <unqualified-name> [<template-args>]
// ::= St <unqualified-name> [<template-args>] # starts with 'S'
//
// Drop and inline:
//
// <type> ::= <operator-name> [<template-args>] # starts with lowercase
// ::= <ctor-dtor-name> [<template-args>] # starts with 'C' or 'D'
// ::= <source-name> [<template-args>] # starts with digit
// ::= <local-source-name> [<template-args>]
// ::= <unnamed-type-name> [<template-args>] # starts with 'U'
//
// One more time:
//
// <type> ::= L <source-name> [<template-args>]
//
// Likewise with <expr-primary>:
//
// <expr-primary> ::= L <type> <expr-cast-value> E
// ::= LZ <encoding> E # cannot overlap; drop
// ::= L <mangled_name> E # cannot overlap; drop
//
// By similar reasoning as shown above, the only <type>s starting with
// <source-name> are "<source-name> [<template-args>]". Inline this.
//
// <expr-primary> ::= L <source-name> [<template-args>] <expr-cast-value> E
//
// Now inline both of these into <template-arg>:
//
// <template-arg> ::= L <source-name> [<template-args>]
// ::= L <source-name> [<template-args>] <expr-cast-value> E
//
// Merge them and we're done:
// <template-arg>
// ::= L <source-name> [<template-args>] [<expr-cast-value> E]
if (ParseLocalSourceName(state) && Optional(ParseTemplateArgs(state))) {
copy = state->parse_state;
if (ParseExprCastValue(state) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return true;
}
// Now that the overlapping cases can't reach this code, we can safely call
// both of these.
if (ParseType(state) || ParseExprPrimary(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'X') && ParseExpression(state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <unresolved-type> ::= <template-param> [<template-args>]
// ::= <decltype>
// ::= <substitution>
static inline bool ParseUnresolvedType(State* state) {
// No ComplexityGuard because we don't copy the state in this stack frame.
return (ParseTemplateParam(state) && Optional(ParseTemplateArgs(state))) ||
ParseDecltype(state) || ParseSubstitution(state, /*accept_std=*/false);
}
// <simple-id> ::= <source-name> [<template-args>]
static inline bool ParseSimpleId(State* state) {
// No ComplexityGuard because we don't copy the state in this stack frame.
// Note: <simple-id> cannot be followed by a parameter pack; see comment in
// ParseUnresolvedType.
return ParseSourceName(state) && Optional(ParseTemplateArgs(state));
}
// <base-unresolved-name> ::= <source-name> [<template-args>]
// ::= on <operator-name> [<template-args>]
// ::= dn <destructor-name>
static bool ParseBaseUnresolvedName(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseSimpleId(state)) {
return true;
}
ParseState copy = state->parse_state;
if (ParseTwoCharToken(state, "on") && ParseOperatorName(state, nullptr) &&
Optional(ParseTemplateArgs(state))) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "dn") &&
(ParseUnresolvedType(state) || ParseSimpleId(state))) {
return true;
}
state->parse_state = copy;
return false;
}
// <unresolved-name> ::= [gs] <base-unresolved-name>
// ::= sr <unresolved-type> <base-unresolved-name>
// ::= srN <unresolved-type> <unresolved-qualifier-level>+ E
// <base-unresolved-name>
// ::= [gs] sr <unresolved-qualifier-level>+ E
// <base-unresolved-name>
static bool ParseUnresolvedName(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (Optional(ParseTwoCharToken(state, "gs")) &&
ParseBaseUnresolvedName(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "sr") && ParseUnresolvedType(state) &&
ParseBaseUnresolvedName(state)) {
return true;
}
state->parse_state = copy;
if (ParseTwoCharToken(state, "sr") && ParseOneCharToken(state, 'N') &&
ParseUnresolvedType(state) &&
OneOrMore(/* <unresolved-qualifier-level> ::= */ ParseSimpleId, state) &&
ParseOneCharToken(state, 'E') && ParseBaseUnresolvedName(state)) {
return true;
}
state->parse_state = copy;
if (Optional(ParseTwoCharToken(state, "gs")) &&
ParseTwoCharToken(state, "sr") &&
OneOrMore(/* <unresolved-qualifier-level> ::= */ ParseSimpleId, state) &&
ParseOneCharToken(state, 'E') && ParseBaseUnresolvedName(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <expression> ::= <1-ary operator-name> <expression>
// ::= <2-ary operator-name> <expression> <expression>
// ::= <3-ary operator-name> <expression> <expression> <expression>
// ::= cl <expression>+ E
// ::= cp <simple-id> <expression>* E # Clang-specific.
// ::= cv <type> <expression> # type (expression)
// ::= cv <type> _ <expression>* E # type (expr-list)
// ::= st <type>
// ::= <template-param>
// ::= <function-param>
// ::= <expr-primary>
// ::= dt <expression> <unresolved-name> # expr.name
// ::= pt <expression> <unresolved-name> # expr->name
// ::= sp <expression> # argument pack expansion
// ::= sr <type> <unqualified-name> <template-args>
// ::= sr <type> <unqualified-name>
// <function-param> ::= fp <(top-level) CV-qualifiers> _
// ::= fp <(top-level) CV-qualifiers> <number> _
// ::= fL <number> p <(top-level) CV-qualifiers> _
// ::= fL <number> p <(top-level) CV-qualifiers> <number> _
static bool ParseExpression(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseTemplateParam(state) || ParseExprPrimary(state)) {
return true;
}
ParseState copy = state->parse_state;
// Object/function call expression.
if (ParseTwoCharToken(state, "cl") && OneOrMore(ParseExpression, state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
// Clang-specific "cp <simple-id> <expression>* E"
// https://clang.llvm.org/doxygen/ItaniumMangle_8cpp_source.html#l04338
if (ParseTwoCharToken(state, "cp") && ParseSimpleId(state) &&
ZeroOrMore(ParseExpression, state) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
// Function-param expression (level 0).
if (ParseTwoCharToken(state, "fp") && Optional(ParseCVQualifiers(state)) &&
Optional(ParseNumber(state, nullptr)) && ParseOneCharToken(state, '_')) {
return true;
}
state->parse_state = copy;
// Function-param expression (level 1+).
if (ParseTwoCharToken(state, "fL") && Optional(ParseNumber(state, nullptr)) &&
ParseOneCharToken(state, 'p') && Optional(ParseCVQualifiers(state)) &&
Optional(ParseNumber(state, nullptr)) && ParseOneCharToken(state, '_')) {
return true;
}
state->parse_state = copy;
// Parse the conversion expressions jointly to avoid re-parsing the <type> in
// their common prefix. Parsed as:
// <expression> ::= cv <type> <conversion-args>
// <conversion-args> ::= _ <expression>* E
// ::= <expression>
//
// Also don't try ParseOperatorName after seeing "cv", since ParseOperatorName
// also needs to accept "cv <type>" in other contexts.
if (ParseTwoCharToken(state, "cv")) {
if (ParseType(state)) {
ParseState copy2 = state->parse_state;
if (ParseOneCharToken(state, '_') && ZeroOrMore(ParseExpression, state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy2;
if (ParseExpression(state)) {
return true;
}
}
} else {
// Parse unary, binary, and ternary operator expressions jointly, taking
// care not to re-parse subexpressions repeatedly. Parse like:
// <expression> ::= <operator-name> <expression>
// [<one-to-two-expressions>]
// <one-to-two-expressions> ::= <expression> [<expression>]
int arity = -1;
if (ParseOperatorName(state, &arity) &&
arity > 0 && // 0 arity => disabled.
(arity < 3 || ParseExpression(state)) &&
(arity < 2 || ParseExpression(state)) &&
(arity < 1 || ParseExpression(state))) {
return true;
}
}
state->parse_state = copy;
// sizeof type
if (ParseTwoCharToken(state, "st") && ParseType(state)) {
return true;
}
state->parse_state = copy;
// Object and pointer member access expressions.
if ((ParseTwoCharToken(state, "dt") || ParseTwoCharToken(state, "pt")) &&
ParseExpression(state) && ParseType(state)) {
return true;
}
state->parse_state = copy;
// Pointer-to-member access expressions. This parses the same as a binary
// operator, but it's implemented separately because "ds" shouldn't be
// accepted in other contexts that parse an operator name.
if (ParseTwoCharToken(state, "ds") && ParseExpression(state) &&
ParseExpression(state)) {
return true;
}
state->parse_state = copy;
// Parameter pack expansion
if (ParseTwoCharToken(state, "sp") && ParseExpression(state)) {
return true;
}
state->parse_state = copy;
return ParseUnresolvedName(state);
}
// <expr-primary> ::= L <type> <(value) number> E
// ::= L <type> <(value) float> E
// ::= L <mangled-name> E
// // A bug in g++'s C++ ABI version 2 (-fabi-version=2).
// ::= LZ <encoding> E
//
// Warning, subtle: the "bug" LZ production above is ambiguous with the first
// production where <type> starts with <local-name>, which can lead to
// exponential backtracking in two scenarios:
//
// - When whatever follows the E in the <local-name> in the first production is
// not a name, we backtrack the whole <encoding> and re-parse the whole thing.
//
// - When whatever follows the <local-name> in the first production is not a
// number and this <expr-primary> may be followed by a name, we backtrack the
// <name> and re-parse it.
//
// Moreover this ambiguity isn't always resolved -- for example, the following
// has two different parses:
//
// _ZaaILZ4aoeuE1x1EvE
// => operator&&<aoeu, x, E, void>
// => operator&&<(aoeu::x)(1), void>
//
// To resolve this, we just do what GCC's demangler does, and refuse to parse
// casts to <local-name> types.
static bool ParseExprPrimary(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
// The "LZ" special case: if we see LZ, we commit to accept "LZ <encoding> E"
// or fail, no backtracking.
if (ParseTwoCharToken(state, "LZ")) {
if (ParseEncoding(state) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// The merged cast production.
if (ParseOneCharToken(state, 'L') && ParseType(state) &&
ParseExprCastValue(state)) {
return true;
}
state->parse_state = copy;
if (ParseOneCharToken(state, 'L') && ParseMangledName(state) &&
ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <number> or <float>, followed by 'E', as described above ParseExprPrimary.
static bool ParseExprCastValue(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
// We have to be able to backtrack after accepting a number because we could
// have e.g. "7fffE", which will accept "7" as a number but then fail to find
// the 'E'.
ParseState copy = state->parse_state;
if (ParseNumber(state, nullptr) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
if (ParseFloatNumber(state) && ParseOneCharToken(state, 'E')) {
return true;
}
state->parse_state = copy;
return false;
}
// <local-name> ::= Z <(function) encoding> E <(entity) name> [<discriminator>]
// ::= Z <(function) encoding> E s [<discriminator>]
//
// Parsing a common prefix of these two productions together avoids an
// exponential blowup of backtracking. Parse like:
// <local-name> := Z <encoding> E <local-name-suffix>
// <local-name-suffix> ::= s [<discriminator>]
// ::= <name> [<discriminator>]
static bool ParseLocalNameSuffix(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (MaybeAppend(state, "::") && ParseName(state) &&
Optional(ParseDiscriminator(state))) {
return true;
}
// Since we're not going to overwrite the above "::" by re-parsing the
// <encoding> (whose trailing '\0' byte was in the byte now holding the
// first ':'), we have to rollback the "::" if the <name> parse failed.
if (state->parse_state.append) {
state->out[state->parse_state.out_cur_idx - 2] = '\0';
}
return ParseOneCharToken(state, 's') && Optional(ParseDiscriminator(state));
}
static bool ParseLocalName(State* state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'Z') && ParseEncoding(state) &&
ParseOneCharToken(state, 'E') && ParseLocalNameSuffix(state)) {
return true;
}
state->parse_state = copy;
return false;
}
// <discriminator> := _ <(non-negative) number>
static bool ParseDiscriminator(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, '_') && ParseNumber(state, nullptr)) {
return true;
}
state->parse_state = copy;
return false;
}
// <substitution> ::= S_
// ::= S <seq-id> _
// ::= St, etc.
//
// "St" is special in that it's not valid as a standalone name, and it *is*
// allowed to precede a name without being wrapped in "N...E". This means that
// if we accept it on its own, we can accept "St1a" and try to parse
// template-args, then fail and backtrack, accept "St" on its own, then "1a" as
// an unqualified name and re-parse the same template-args. To block this
// exponential backtracking, we disable it with 'accept_std=false' in
// problematic contexts.
static bool ParseSubstitution(State* state, bool accept_std) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseTwoCharToken(state, "S_")) {
MaybeAppend(state, "?"); // We don't support substitutions.
return true;
}
ParseState copy = state->parse_state;
if (ParseOneCharToken(state, 'S') && ParseSeqId(state) &&
ParseOneCharToken(state, '_')) {
MaybeAppend(state, "?"); // We don't support substitutions.
return true;
}
state->parse_state = copy;
// Expand abbreviations like "St" => "std".
if (ParseOneCharToken(state, 'S')) {
const AbbrevPair *p;
for (p = kSubstitutionList; p->abbrev != nullptr; ++p) {
if (RemainingInput(state)[0] == p->abbrev[1] &&
(accept_std || p->abbrev[1] != 't')) {
MaybeAppend(state, "std");
if (p->real_name[0] != '\0') {
MaybeAppend(state, "::");
MaybeAppend(state, p->real_name);
}
++state->parse_state.mangled_idx;
return true;
}
}
}
state->parse_state = copy;
return false;
}
// Parse <mangled-name>, optionally followed by either a function-clone suffix
// or version suffix. Returns true only if all of "mangled_cur" was consumed.
static bool ParseTopLevelMangledName(State *state) {
ComplexityGuard guard(state);
if (guard.IsTooComplex()) {
return false;
}
if (ParseMangledName(state)) {
if (RemainingInput(state)[0] != '\0') {
// Drop trailing function clone suffix, if any.
if (IsFunctionCloneSuffix(RemainingInput(state))) {
return true;
}
// Append trailing version suffix if any.
// ex. _Z3foo@@GLIBCXX_3.4
if (RemainingInput(state)[0] == '@') {
MaybeAppend(state, RemainingInput(state));
return true;
}
return false; // Unconsumed suffix.
}
return true;
}
return false;
}
static bool Overflowed(const State* state) {
return state->parse_state.out_cur_idx >= state->out_end_idx;
}
#endif
// The demangler entry point.
bool Demangle(const char* mangled, char* out, size_t out_size) {
#if defined(GLOG_OS_WINDOWS)
#if defined(HAVE_DBGHELP)
// When built with incremental linking, the Windows debugger
// library provides a more complicated `Symbol->Name` with the
// Incremental Linking Table offset, which looks like
// `@ILT+1105(?func@Foo@@SAXH@Z)`. However, the demangler expects
// only the mangled symbol, `?func@Foo@@SAXH@Z`. Fortunately, the
// mangled symbol is guaranteed not to have parentheses,
// so we search for `(` and extract up to `)`.
//
// Since we may be in a signal handler here, we cannot use `std::string`.
char buffer[1024]; // Big enough for a sane symbol.
const char *lparen = strchr(mangled, '(');
if (lparen) {
// Extract the string `(?...)`
const char *rparen = strchr(lparen, ')');
size_t length = static_cast<size_t>(rparen - lparen) - 1;
strncpy(buffer, lparen + 1, length);
buffer[length] = '\0';
mangled = buffer;
} // Else the symbol wasn't inside a set of parentheses
// We use the ANSI version to ensure the string type is always `char *`.
return UnDecorateSymbolName(mangled, out, out_size, UNDNAME_COMPLETE);
#else
(void)mangled;
(void)out;
(void)out_size;
return false;
#endif
#else
State state;
InitState(&state, mangled, out, out_size);
return ParseTopLevelMangledName(&state) && !Overflowed(&state) &&
state.parse_state.out_cur_idx > 0;
#endif
}
_END_GOOGLE_NAMESPACE_
|