https://github.com/python/cpython/commit/451ece914bf61260863da8349eb0ea29dddb75ba commit: 451ece914bf61260863da8349eb0ea29dddb75ba branch: 3.10 author: Stan Ulbrych <[email protected]> committer: pablogsal <[email protected]> date: 2026-08-13T01:29:50+01:00 summary:
[3.10] gh-155558: Update bundled Expat to 2.8.3 (GH-155560) (#155618) * [3.10] gh-155558: Update bundled Expat to 2.8.3 (GH-155560) (cherry picked from commit 3d6a50598ba7c89a6468f8e77afe2a39787c80e5) Co-authored-by: Stan Ulbrych <[email protected]> * Don't use role unsupported by old Sphnx files: A Misc/NEWS.d/next/Security/2026-08-11-15-24-21.gh-issue-155558.yDsXZC.rst M Modules/expat/ascii.h M Modules/expat/asciitab.h M Modules/expat/expat.h M Modules/expat/expat_external.h M Modules/expat/fallthrough.h M Modules/expat/iasciitab.h M Modules/expat/internal.h M Modules/expat/latin1tab.h M Modules/expat/memory_sanitizer.h M Modules/expat/nametab.h M Modules/expat/refresh.sh M Modules/expat/siphash.h M Modules/expat/utf8tab.h M Modules/expat/winconfig.h M Modules/expat/xcsinc.c M Modules/expat/xmlparse.c M Modules/expat/xmlrole.c M Modules/expat/xmlrole.h M Modules/expat/xmltok.c M Modules/expat/xmltok.h M Modules/expat/xmltok_impl.c M Modules/expat/xmltok_impl.h M Modules/expat/xmltok_ns.c diff --git a/Misc/NEWS.d/next/Security/2026-08-11-15-24-21.gh-issue-155558.yDsXZC.rst b/Misc/NEWS.d/next/Security/2026-08-11-15-24-21.gh-issue-155558.yDsXZC.rst new file mode 100644 index 000000000000000..90cc3a765cf554e --- /dev/null +++ b/Misc/NEWS.d/next/Security/2026-08-11-15-24-21.gh-issue-155558.yDsXZC.rst @@ -0,0 +1,2 @@ +Update bundled `libexpat <https://libexpat.github.io/>`_ to version 2.8.3 +for the fix to `CVE-2026-72522 <https://www.cve.org/CVERecord?id=CVE-2026-72522>`_. diff --git a/Modules/expat/ascii.h b/Modules/expat/ascii.h index 1f594d2e54b4d29..1d9cf70bf695da1 100644 --- a/Modules/expat/ascii.h +++ b/Modules/expat/ascii.h @@ -31,6 +31,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #define ASCII_A 0x41 diff --git a/Modules/expat/asciitab.h b/Modules/expat/asciitab.h index af766fb24785ea3..43af0876ccd3f8a 100644 --- a/Modules/expat/asciitab.h +++ b/Modules/expat/asciitab.h @@ -30,6 +30,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ /* 0x00 */ BT_NONXML, BT_NONXML, BT_NONXML, BT_NONXML, diff --git a/Modules/expat/expat.h b/Modules/expat/expat.h index c493c70441c4184..dbebd985a652ac7 100644 --- a/Modules/expat/expat.h +++ b/Modules/expat/expat.h @@ -40,6 +40,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef Expat_INCLUDED @@ -1094,7 +1096,7 @@ XML_SetReparseDeferralEnabled(XML_Parser parser, XML_Bool enabled); */ # define XML_MAJOR_VERSION 2 # define XML_MINOR_VERSION 8 -# define XML_MICRO_VERSION 2 +# define XML_MICRO_VERSION 3 # ifdef __cplusplus } diff --git a/Modules/expat/expat_external.h b/Modules/expat/expat_external.h index cc945c424e471f7..6cc3f19ed7dbc89 100644 --- a/Modules/expat/expat_external.h +++ b/Modules/expat/expat_external.h @@ -36,6 +36,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef Expat_External_INCLUDED diff --git a/Modules/expat/fallthrough.h b/Modules/expat/fallthrough.h index 707dbdd44bfe53d..0152d1bdc5de8ff 100644 --- a/Modules/expat/fallthrough.h +++ b/Modules/expat/fallthrough.h @@ -27,6 +27,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef FALLTHROUGH_H diff --git a/Modules/expat/iasciitab.h b/Modules/expat/iasciitab.h index 5d8646f2a318b8a..1de8d519c8b7bad 100644 --- a/Modules/expat/iasciitab.h +++ b/Modules/expat/iasciitab.h @@ -30,6 +30,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ /* Like asciitab.h, except that 0xD has code BT_S rather than BT_CR */ diff --git a/Modules/expat/internal.h b/Modules/expat/internal.h index 420d4217a569b1d..7e67d2e378c5243 100644 --- a/Modules/expat/internal.h +++ b/Modules/expat/internal.h @@ -53,6 +53,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #if defined(__GNUC__) && defined(__i386__) && ! defined(__MINGW32__) @@ -123,20 +125,11 @@ # define EXPAT_FMT_SIZE_T(midpart) "%" midpart "u" # endif #else +# include <inttypes.h> // PRIdPTR, PRIuPTR # define EXPAT_FMT_LLX(midpart) "%" midpart "llx" # define EXPAT_FMT_ULL(midpart) "%" midpart "llu" -# if ! defined(ULONG_MAX) -# error Compiler did not define ULONG_MAX for us -# elif ULONG_MAX == 18446744073709551615u // 2^64-1 -# define EXPAT_FMT_PTRDIFF_T(midpart) "%" midpart "ld" -# define EXPAT_FMT_SIZE_T(midpart) "%" midpart "lu" -# elif defined(__wasm32__) // 32bit mode Emscripten or WASI SDK -# define EXPAT_FMT_PTRDIFF_T(midpart) "%" midpart "ld" -# define EXPAT_FMT_SIZE_T(midpart) "%" midpart "zu" -# else -# define EXPAT_FMT_PTRDIFF_T(midpart) "%" midpart "d" -# define EXPAT_FMT_SIZE_T(midpart) "%" midpart "u" -# endif +# define EXPAT_FMT_PTRDIFF_T(midpart) "%" midpart PRIdPTR +# define EXPAT_FMT_SIZE_T(midpart) "%" midpart PRIuPTR #endif #ifndef UNUSED_P diff --git a/Modules/expat/latin1tab.h b/Modules/expat/latin1tab.h index b681d278af6569b..3793f4f3cc7841d 100644 --- a/Modules/expat/latin1tab.h +++ b/Modules/expat/latin1tab.h @@ -30,6 +30,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ /* 0x80 */ BT_OTHER, BT_OTHER, BT_OTHER, BT_OTHER, diff --git a/Modules/expat/memory_sanitizer.h b/Modules/expat/memory_sanitizer.h index a8a8006ccded12d..f739b88f304871b 100644 --- a/Modules/expat/memory_sanitizer.h +++ b/Modules/expat/memory_sanitizer.h @@ -27,6 +27,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #if ! defined(MEMORY_SANITIZER_H) diff --git a/Modules/expat/nametab.h b/Modules/expat/nametab.h index 63485446b967272..2385851562e7dcd 100644 --- a/Modules/expat/nametab.h +++ b/Modules/expat/nametab.h @@ -28,6 +28,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ static const unsigned namingBitmap[] = { diff --git a/Modules/expat/refresh.sh b/Modules/expat/refresh.sh index 8d476ac2fbf24a7..66fb0a696e8eff1 100755 --- a/Modules/expat/refresh.sh +++ b/Modules/expat/refresh.sh @@ -12,9 +12,9 @@ fi # Update this when updating to a new version after verifying that the changes # the update brings in are good. These values are used for verifying the SBOM, too. -expected_libexpat_tag="R_2_8_2" -expected_libexpat_version="2.8.2" -expected_libexpat_sha256="ef7d1994f533c9e7343d6c19f31064fc8ebbcbcaa144be3812b4f43052a05f4c" +expected_libexpat_tag="R_2_8_3" +expected_libexpat_version="2.8.3" +expected_libexpat_sha256="22920a86c83f32300b11463635b71f11137a917975af297725e55525027d4e50" expat_dir="$(realpath "$(dirname -- "${BASH_SOURCE[0]}")")" cd ${expat_dir} diff --git a/Modules/expat/siphash.h b/Modules/expat/siphash.h index be216b4006dc847..ac1fbfaf0bde81d 100644 --- a/Modules/expat/siphash.h +++ b/Modules/expat/siphash.h @@ -8,6 +8,8 @@ * * 1. https://www.131002.net/siphash/siphash24.c * 2. https://www.131002.net/siphash/ + * + * SPDX-License-Identifier: CC0-1.0 * -------------------------------------------------------------------------- * HISTORY: * diff --git a/Modules/expat/utf8tab.h b/Modules/expat/utf8tab.h index 88efcf91cc16a6c..73732d1527f1420 100644 --- a/Modules/expat/utf8tab.h +++ b/Modules/expat/utf8tab.h @@ -30,6 +30,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ /* 0x80 */ BT_TRAIL, BT_TRAIL, BT_TRAIL, BT_TRAIL, diff --git a/Modules/expat/winconfig.h b/Modules/expat/winconfig.h index 05805514ec7fa21..9e7e8bdc22ca36c 100644 --- a/Modules/expat/winconfig.h +++ b/Modules/expat/winconfig.h @@ -31,6 +31,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef WINCONFIG_H diff --git a/Modules/expat/xcsinc.c b/Modules/expat/xcsinc.c index 3597c2480bc9617..675b2844c708d6a 100644 --- a/Modules/expat/xcsinc.c +++ b/Modules/expat/xcsinc.c @@ -6,7 +6,7 @@ \___/_/\_\ .__/ \__,_|\__| |_| XML parser - Copyright (c) 2022 Sebastian Pipping <[email protected]> + Copyright (c) 2022-2026 Sebastian Pipping <[email protected]> Licensed under the MIT license: Permission is hereby granted, free of charge, to any person obtaining @@ -27,8 +27,14 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ +#if defined(XML_UNICODE) && defined(XML_UNICODE_WCHAR_T) +# include <wchar.h> +#endif + static size_t xcslen(const XML_Char *s) { #ifdef XML_UNICODE diff --git a/Modules/expat/xmlparse.c b/Modules/expat/xmlparse.c index d961ebf227e7405..4fa61bca8c16293 100644 --- a/Modules/expat/xmlparse.c +++ b/Modules/expat/xmlparse.c @@ -1,4 +1,4 @@ -/* 5de44e6750c6cc78818f06ed552f522a1241df0299395250e1792cb339389daf (2.8.2+) +/* ee5f82c3ffd57c5224394ba46f348dbce466d34d6c925a527ae46b1cfe6adf1d (2.8.3+) __ __ _ ___\ \/ /_ __ __ _| |_ / _ \\ /| '_ \ / _` | __| @@ -50,6 +50,7 @@ Copyright (c) 2026 Nick Begg <[email protected]> Copyright (c) 2026 Kartik Kenchi <[email protected]> Copyright (c) 2026 Haris Hussain <[email protected]> + Copyright (c) 2026 Evgeny Kotkov <[email protected]> Licensed under the MIT license: Permission is hereby granted, free of charge, to any person obtaining @@ -70,6 +71,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #define XML_BUILDING_EXPAT 1 @@ -93,10 +96,10 @@ #include <stddef.h> #include <string.h> /* memset(), memcpy() */ #include <assert.h> -#include <limits.h> /* INT_MAX, LLONG_MAX, LONG_MAX, UINT_MAX */ +#include <limits.h> /* INT_MAX, UINT_MAX */ #include <stdio.h> /* fprintf */ #include <stdlib.h> /* getenv */ -#include <stdint.h> /* SIZE_MAX, uintptr_t */ +#include <stdint.h> /* SIZE_MAX, UINT64_MAX, uint64_t, uintptr_t */ #include <math.h> /* isnan */ #include <errno.h> @@ -211,12 +214,6 @@ typedef char ICHAR; #endif -#ifdef XML_LARGE_SIZE -# define XML_INDEX_MAX LLONG_MAX -#else -# define XML_INDEX_MAX LONG_MAX -#endif - /* Round up n to be a multiple of sz, where sz is a power of 2. */ #define ROUND_UP(n, sz) (((n) + ((sz) - 1)) & ~((sz) - 1)) @@ -719,7 +716,7 @@ struct XML_ParserStruct { char *m_bufferEnd; // past last character to be parsed const char *m_bufferLim; // allocated end of m_buffer - XML_Index m_parseEndByteIndex; + uint64_t m_parseEndByteIndex; const char *m_parseEndPtr; size_t m_partialTokenBytesBefore; /* used in heuristic to avoid O(n^2) */ XML_Bool m_reparseDeferralEnabled; @@ -2312,7 +2309,7 @@ XML_Parse(XML_Parser parser, const char *s, int len, int isFinal) { int nLeftOver; enum XML_Status result; /* Detect overflow (a+b > MAX <==> b > MAX-a) */ - if (len > XML_INDEX_MAX - parser->m_parseEndByteIndex) { + if ((uint64_t)len > UINT64_MAX - parser->m_parseEndByteIndex) { parser->m_errorCode = XML_ERROR_NO_MEMORY; parser->m_eventPtr = parser->m_eventEndPtr = NULL; parser->m_processor = errorProcessor; @@ -2430,7 +2427,7 @@ XML_ParseBuffer(XML_Parser parser, int len, int isFinal) { } // Detect and avoid integer overflow - if (len > XML_INDEX_MAX - parser->m_parseEndByteIndex) { + if ((uint64_t)len > UINT64_MAX - parser->m_parseEndByteIndex) { parser->m_errorCode = XML_ERROR_NO_MEMORY; parser->m_eventPtr = parser->m_eventEndPtr = NULL; parser->m_processor = errorProcessor; @@ -2692,9 +2689,15 @@ XML_Index XMLCALL XML_GetCurrentByteIndex(XML_Parser parser) { if (parser == NULL) return -1; - if (parser->m_eventPtr) + if (parser->m_eventPtr) { + // NOTE: XML_Index is known to wrap around for >2 GiB content + // on 32bit machines and 64bit Windows, unless (non-default and + // uncommon) XML_LARGE_SIZE is defined. + // That's a bug and it only lives on because we cannot break + // ABI compatibility of public API. return (XML_Index)(parser->m_parseEndByteIndex - (parser->m_parseEndPtr - parser->m_eventPtr)); + } return -1; } @@ -2736,7 +2739,12 @@ XML_GetCurrentLineNumber(XML_Parser parser) { parser->m_eventPtr, &parser->m_position); parser->m_positionPtr = parser->m_eventPtr; } - return parser->m_position.lineNumber + 1; + // NOTE: XML_Size is known to wrap around for >2 4iB content + // on 32bit machines and 64bit Windows, unless (non-default and + // uncommon) XML_LARGE_SIZE is defined. + // That's a bug and it only lives on because we cannot break + // ABI compatibility of public API. + return (XML_Size)(parser->m_position.lineNumber + 1); } XML_Size XMLCALL @@ -2748,7 +2756,12 @@ XML_GetCurrentColumnNumber(XML_Parser parser) { parser->m_eventPtr, &parser->m_position); parser->m_positionPtr = parser->m_eventPtr; } - return parser->m_position.columnNumber; + // NOTE: XML_Size is known to wrap around for >2 4iB content + // on 32bit machines and 64bit Windows, unless (non-default and + // uncommon) XML_LARGE_SIZE is defined. + // That's a bug and it only lives on because we cannot break + // ABI compatibility of public API. + return (XML_Size)parser->m_position.columnNumber; } void XMLCALL @@ -3905,14 +3918,22 @@ storeAtts(XML_Parser parser, const ENCODING *enc, const char *attStr, if (! attId) return XML_ERROR_NO_MEMORY; #ifdef XML_ATTR_INFO + // NOTE: XML_Index is known to wrap around for >2 GiB content + // on 32bit machines and 64bit Windows, unless (non-default and + // uncommon) XML_LARGE_SIZE is defined. + // That's a bug and it only lives on because we cannot break + // ABI compatibility of public API. currAttInfo->nameStart - = parser->m_parseEndByteIndex - (parser->m_parseEndPtr - currAtt->name); + = (XML_Index)(parser->m_parseEndByteIndex + - (parser->m_parseEndPtr - currAtt->name)); currAttInfo->nameEnd = currAttInfo->nameStart + XmlNameLength(enc, currAtt->name); - currAttInfo->valueStart = parser->m_parseEndByteIndex - - (parser->m_parseEndPtr - currAtt->valuePtr); - currAttInfo->valueEnd = parser->m_parseEndByteIndex - - (parser->m_parseEndPtr - currAtt->valueEnd); + currAttInfo->valueStart + = (XML_Index)(parser->m_parseEndByteIndex + - (parser->m_parseEndPtr - currAtt->valuePtr)); + currAttInfo->valueEnd + = (XML_Index)(parser->m_parseEndByteIndex + - (parser->m_parseEndPtr - currAtt->valueEnd)); #endif /* Detect duplicate attributes by their QNames. This does not work when namespace processing is turned on and different prefixes for the same @@ -6554,11 +6575,12 @@ storeAttributeValue(XML_Parser parser, const ENCODING *enc, XML_Bool isCdata, // Check if entity is complete, if not, mark down how much of it is // processed. A XML_SUSPENDED check here is not required as // appendAttributeValue will never suspend the parser. - if (textEnd != nextInEntity) { + if (nextInEntity < textEnd) { entity->processed = (int)(nextInEntity - (const char *)entity->textPtr); continue; } + assert(nextInEntity == textEnd); // Entity is complete. We cannot close it here since we need to first // process its possible inner entities (which are added to the @@ -8192,7 +8214,7 @@ poolGrow(STRING_POOL *pool) { pool->freeBlocks = tem; memcpy(pool->blocks->s, pool->start, (pool->end - pool->start) * sizeof(XML_Char)); - pool->ptr = pool->blocks->s + (pool->ptr - pool->start); + pool->ptr = pool->blocks->s + EXPAT_SAFE_PTR_DIFF(pool->ptr, pool->start); pool->start = pool->blocks->s; pool->end = pool->start + pool->blocks->size; return XML_TRUE; @@ -8205,7 +8227,8 @@ poolGrow(STRING_POOL *pool) { /* NOTE: Needs to be calculated prior to calling `realloc` to avoid dangling pointers: */ - const ptrdiff_t offsetInsideBlock = pool->ptr - pool->start; + const ptrdiff_t offsetInsideBlock + = EXPAT_SAFE_PTR_DIFF(pool->ptr, pool->start); if (blockSize < 0) { /* This condition traps a situation where either more than @@ -8268,8 +8291,9 @@ poolGrow(STRING_POOL *pool) { tem->next = pool->blocks; pool->blocks = tem; if (pool->ptr != pool->start) - memcpy(tem->s, pool->start, (pool->ptr - pool->start) * sizeof(XML_Char)); - pool->ptr = tem->s + (pool->ptr - pool->start); + memcpy(tem->s, pool->start, + EXPAT_SAFE_PTR_DIFF(pool->ptr, pool->start) * sizeof(XML_Char)); + pool->ptr = tem->s + EXPAT_SAFE_PTR_DIFF(pool->ptr, pool->start); pool->start = tem->s; pool->end = tem->s + blockSize; } diff --git a/Modules/expat/xmlrole.c b/Modules/expat/xmlrole.c index d56bee82dd2d13f..8f6bedda71652d7 100644 --- a/Modules/expat/xmlrole.c +++ b/Modules/expat/xmlrole.c @@ -37,6 +37,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #include "expat_config.h" diff --git a/Modules/expat/xmlrole.h b/Modules/expat/xmlrole.h index 9d0d4ff11b7f982..903a6951f434e2e 100644 --- a/Modules/expat/xmlrole.h +++ b/Modules/expat/xmlrole.h @@ -31,6 +31,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef XmlRole_INCLUDED diff --git a/Modules/expat/xmltok.c b/Modules/expat/xmltok.c index 387c6e44a25f80c..d66af6045f07aca 100644 --- a/Modules/expat/xmltok.c +++ b/Modules/expat/xmltok.c @@ -12,7 +12,7 @@ Copyright (c) 2002 Greg Stein <[email protected]> Copyright (c) 2002-2016 Karl Waclawek <[email protected]> Copyright (c) 2005-2009 Steven Solie <[email protected]> - Copyright (c) 2016-2024 Sebastian Pipping <[email protected]> + Copyright (c) 2016-2026 Sebastian Pipping <[email protected]> Copyright (c) 2016 Pascal Cuoq <[email protected]> Copyright (c) 2016 Don Lewis <[email protected]> Copyright (c) 2017 Rhodri James <[email protected]> @@ -26,6 +26,7 @@ Copyright (c) 2023 Hanno Böck <[email protected]> Copyright (c) 2025 Alfonso Gregory <[email protected]> Copyright (c) 2026 Nick Begg <[email protected]> + Copyright (c) 2026 Kartik Kenchi <[email protected]> Licensed under the MIT license: Permission is hereby granted, free of charge, to any person obtaining @@ -46,6 +47,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #include "expat_config.h" @@ -705,9 +708,10 @@ unicode_byte_type(char hi, char lo) { enum XML_Convert_Result res = XML_CONVERT_COMPLETED; \ UNUSED_P(enc); \ fromLim = *fromP + (((fromLim - *fromP) >> 1) << 1); /* shrink to even */ \ - /* Avoid copying first half only of surrogate */ \ + /* Avoid copying the first half (2 bytes) of surrogate pairs (4 bytes) */ \ if (fromLim - *fromP > ((toLim - *toP) << 1) \ - && (GET_HI(fromLim - 2) & 0xF8) == 0xD8) { \ + && /* are the last two bytes a high surrogate (0xD800-0xDBFF)? */ \ + (GET_HI(fromLim - 2) & 0xFC) == 0xD8) { \ fromLim -= 2; \ res = XML_CONVERT_INPUT_INCOMPLETE; \ } \ @@ -1177,6 +1181,13 @@ doParseXmlDecl(const ENCODING *(*encodingFinder)(const ENCODING *, const char *, *versionPtr = val; if (versionEndPtr) *versionEndPtr = ptr; + /* The version number must not be empty; VersionNum requires at least + one character. The encoding and standalone pseudo-attributes below + already reject an empty value, so keep version consistent. */ + if (val == ptr - enc->minBytesPerChar) { + *badPtr = val; + return 0; + } if (! parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) { *badPtr = ptr; return 0; diff --git a/Modules/expat/xmltok.h b/Modules/expat/xmltok.h index 79a9fb76871f10b..bd868b87a407d67 100644 --- a/Modules/expat/xmltok.h +++ b/Modules/expat/xmltok.h @@ -10,7 +10,7 @@ Copyright (c) 2000 Clark Cooper <[email protected]> Copyright (c) 2002 Fred L. Drake, Jr. <[email protected]> Copyright (c) 2002-2005 Karl Waclawek <[email protected]> - Copyright (c) 2016-2024 Sebastian Pipping <[email protected]> + Copyright (c) 2016-2026 Sebastian Pipping <[email protected]> Copyright (c) 2017 Rhodri James <[email protected]> Licensed under the MIT license: @@ -32,11 +32,15 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifndef XmlTok_INCLUDED # define XmlTok_INCLUDED 1 +# include <stdint.h> // uint64_t + # ifdef __cplusplus extern "C" { # endif @@ -145,8 +149,8 @@ extern "C" { typedef struct position { /* first line and first column are 0 not 1 */ - XML_Size lineNumber; - XML_Size columnNumber; + uint64_t lineNumber; + uint64_t columnNumber; } POSITION; typedef struct { diff --git a/Modules/expat/xmltok_impl.c b/Modules/expat/xmltok_impl.c index eae11bdd968aee4..1d8e457d9de5445 100644 --- a/Modules/expat/xmltok_impl.c +++ b/Modules/expat/xmltok_impl.c @@ -38,6 +38,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifdef XML_TOK_IMPL_C diff --git a/Modules/expat/xmltok_impl.h b/Modules/expat/xmltok_impl.h index 3469c4ae138c95a..13a08899da3b1a5 100644 --- a/Modules/expat/xmltok_impl.h +++ b/Modules/expat/xmltok_impl.h @@ -29,6 +29,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ enum { diff --git a/Modules/expat/xmltok_ns.c b/Modules/expat/xmltok_ns.c index 810ca2c6d0485eb..40a02a0308b563c 100644 --- a/Modules/expat/xmltok_ns.c +++ b/Modules/expat/xmltok_ns.c @@ -33,6 +33,8 @@ DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + + SPDX-License-Identifier: MIT */ #ifdef XML_TOK_NS_C _______________________________________________ Python-checkins mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3//lists/python-checkins.python.org Member address: [email protected]
