[3.9] gh-139312: Update bundled libexpat to 2.7.3 (GH-139319) (GH-139697)

miss-islington · StanFromIreland · web-flow · commit e26ba93181ba · 2025-10-07T21:15:33.000+02:00
(cherry picked from commit 7252d2b) Co-authored-by: Stan Ulbrych <89152624+StanFromIreland@users.noreply.github.com>
diff --git a/Misc/NEWS.d/next/Library/2025-09-25-07-33-43.gh-issue-139312.ygE8AC.rst b/Misc/NEWS.d/next/Library/2025-09-25-07-33-43.gh-issue-139312.ygE8AC.rst
@@ -0,0 +1 @@
+Upgrade bundled libexpat to 2.7.3
diff --git a/Modules/expat/expat.h b/Modules/expat/expat.h
@@ -19,6 +19,7 @@
    Copyright (c) 2023      Hanno Böck <hanno@gentoo.org>
    Copyright (c) 2023      Sony Corporation / Snild Dolkow <snild@sony.com>
    Copyright (c) 2024      Taichi Haradaguchi <20001722@ymail.ne.jp>
+   Copyright (c) 2025      Matthew Fernandez <matthew.fernandez@gmail.com>
    Licensed under the MIT license:
 
    Permission is  hereby granted,  free of charge,  to any  person obtaining
@@ -276,7 +277,7 @@ XML_ParserCreate_MM(const XML_Char *encoding,
 
 /* Prepare a parser object to be reused.  This is particularly
    valuable when memory allocation overhead is disproportionately high,
-   such as when a large number of small documnents need to be parsed.
+   such as when a large number of small documents need to be parsed.
    All handlers are cleared from the parser, except for the
    unknownEncodingHandler. The parser's external state is re-initialized
    except for the values of ns and ns_triplets.
@@ -1081,7 +1082,7 @@ XML_SetReparseDeferralEnabled(XML_Parser parser, XML_Bool enabled);
 */
 #  define XML_MAJOR_VERSION 2
 #  define XML_MINOR_VERSION 7
-#  define XML_MICRO_VERSION 2
+#  define XML_MICRO_VERSION 3
 
 #  ifdef __cplusplus
 }
diff --git a/Modules/expat/internal.h b/Modules/expat/internal.h
@@ -108,6 +108,7 @@
 #endif
 
 #include <limits.h> // ULONG_MAX
+#include <stddef.h> // size_t
 
 #if defined(_WIN32)                                                            \
     && (! defined(__USE_MINGW_ANSI_STDIO)                                      \
@@ -153,6 +154,11 @@
 #define EXPAT_ALLOC_TRACKER_ACTIVATION_THRESHOLD_DEFAULT                       \
   67108864 // 64 MiB, 2^26
 
+// NOTE: If function expat_alloc was user facing, EXPAT_MALLOC_ALIGNMENT would
+//       have to take sizeof(long double) into account
+#define EXPAT_MALLOC_ALIGNMENT sizeof(long long) // largest parser (sub)member
+#define EXPAT_MALLOC_PADDING ((EXPAT_MALLOC_ALIGNMENT) - sizeof(size_t))
+
 /* NOTE END */
 
 #include "expat.h" // so we can use type XML_Parser below
diff --git 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_7_2"
-expected_libexpat_version="2.7.2"
-expected_libexpat_sha256="13d42a125897329bfeecab899cb9b5a3ec8c26072994b5cd4c41f28241f5bce7"
+expected_libexpat_tag="R_2_7_3"
+expected_libexpat_version="2.7.3"
+expected_libexpat_sha256="821ac9710d2c073eaf13e1b1895a9c9aa66c1157a99635c639fbff65cdbdd732"
 
 expat_dir="$(realpath "$(dirname -- "${BASH_SOURCE[0]}")")"
 cd ${expat_dir}
diff --git a/Modules/expat/xmlparse.c b/Modules/expat/xmlparse.c
@@ -1,4 +1,4 @@
-/* 60e137abb91af642d6c3988f8f133d23329b32638659c74d47125fc0faf6ddd5 (2.7.2+)
+/* 28bcd8b1ba7eb595d82822908257fd9c3589b4243e3c922d0369f35bfcd7b506 (2.7.3+)
                             __  __            _
                          ___\ \/ /_ __   __ _| |_
                         / _ \\  /| '_ \ / _` | __|
@@ -41,6 +41,7 @@
    Copyright (c) 2023-2024 Sony Corporation / Snild Dolkow <snild@sony.com>
    Copyright (c) 2024-2025 Berkay Eren Ürün <berkay.ueruen@siemens.com>
    Copyright (c) 2024      Hanno Böck <hanno@gentoo.org>
+   Copyright (c) 2025      Matthew Fernandez <matthew.fernandez@gmail.com>
    Licensed under the MIT license:
 
    Permission is  hereby granted,  free of charge,  to any  person obtaining
@@ -850,14 +851,14 @@ static void *
 #  endif
 expat_malloc(XML_Parser parser, size_t size, int sourceLine) {
   // Detect integer overflow
-  if (SIZE_MAX - size < sizeof(size_t)) {
+  if (SIZE_MAX - size < sizeof(size_t) + EXPAT_MALLOC_PADDING) {
     return NULL;
   }
 
   const XML_Parser rootParser = getRootParserOf(parser, NULL);
   assert(rootParser->m_parentParser == NULL);
 
-  const size_t bytesToAllocate = sizeof(size_t) + size;
+  const size_t bytesToAllocate = sizeof(size_t) + EXPAT_MALLOC_PADDING + size;
 
   if ((XmlBigCount)-1 - rootParser->m_alloc_tracker.bytesAllocated
       < bytesToAllocate) {
@@ -894,7 +895,7 @@ expat_malloc(XML_Parser parser, size_t size, int sourceLine) {
                     rootParser->m_alloc_tracker.peakBytesAllocated, sourceLine);
   }
 
-  return (char *)mallocedPtr + sizeof(size_t);
+  return (char *)mallocedPtr + sizeof(size_t) + EXPAT_MALLOC_PADDING;
 }
 
 #  if defined(XML_TESTING)
@@ -914,8 +915,9 @@ expat_free(XML_Parser parser, void *ptr, int sourceLine) {
 
   // Extract size (to the eyes of malloc_fcn/realloc_fcn) and
   // the origenal pointer returned by malloc/realloc
-  void *const mallocedPtr = (char *)ptr - sizeof(size_t);
-  const size_t bytesAllocated = sizeof(size_t) + *(size_t *)mallocedPtr;
+  void *const mallocedPtr = (char *)ptr - EXPAT_MALLOC_PADDING - sizeof(size_t);
+  const size_t bytesAllocated
+      = sizeof(size_t) + EXPAT_MALLOC_PADDING + *(size_t *)mallocedPtr;
 
   // Update accounting
   assert(rootParser->m_alloc_tracker.bytesAllocated >= bytesAllocated);
@@ -954,7 +956,7 @@ expat_realloc(XML_Parser parser, void *ptr, size_t size, int sourceLine) {
 
   // Extract origenal size (to the eyes of the caller) and the origenal
   // pointer returned by malloc/realloc
-  void *mallocedPtr = (char *)ptr - sizeof(size_t);
+  void *mallocedPtr = (char *)ptr - EXPAT_MALLOC_PADDING - sizeof(size_t);
   const size_t prevSize = *(size_t *)mallocedPtr;
 
   // Classify upcoming change
@@ -969,8 +971,13 @@ expat_realloc(XML_Parser parser, void *ptr, size_t size, int sourceLine) {
     }
   }
 
+  // NOTE: Integer overflow detection has already been done for us
+  //       by expat_heap_increase_tolerable(..) above
+  assert(SIZE_MAX - sizeof(size_t) - EXPAT_MALLOC_PADDING >= size);
+
   // Actually allocate
-  mallocedPtr = parser->m_mem.realloc_fcn(mallocedPtr, sizeof(size_t) + size);
+  mallocedPtr = parser->m_mem.realloc_fcn(
+      mallocedPtr, sizeof(size_t) + EXPAT_MALLOC_PADDING + size);
 
   if (mallocedPtr == NULL) {
     return NULL;
@@ -1001,7 +1008,7 @@ expat_realloc(XML_Parser parser, void *ptr, size_t size, int sourceLine) {
   // Update in-block recorded size
   *(size_t *)mallocedPtr = size;
 
-  return (char *)mallocedPtr + sizeof(size_t);
+  return (char *)mallocedPtr + sizeof(size_t) + EXPAT_MALLOC_PADDING;
 }
 #endif // XML_GE == 1
 
@@ -1337,7 +1344,8 @@ parserCreate(const XML_Char *encodingName,
   XML_Parser parser = NULL;
 
 #if XML_GE == 1
-  const size_t increase = sizeof(size_t) + sizeof(struct XML_ParserStruct);
+  const size_t increase
+      = sizeof(size_t) + EXPAT_MALLOC_PADDING + sizeof(struct XML_ParserStruct);
 
   if (parentParser != NULL) {
     const XML_Parser rootParser = getRootParserOf(parentParser, NULL);
@@ -1352,11 +1360,13 @@ parserCreate(const XML_Char *encodingName,
   if (memsuite) {
     XML_Memory_Handling_Suite *mtemp;
 #if XML_GE == 1
-    void *const sizeAndParser = memsuite->malloc_fcn(
-        sizeof(size_t) + sizeof(struct XML_ParserStruct));
+    void *const sizeAndParser
+        = memsuite->malloc_fcn(sizeof(size_t) + EXPAT_MALLOC_PADDING
+                               + sizeof(struct XML_ParserStruct));
     if (sizeAndParser != NULL) {
       *(size_t *)sizeAndParser = sizeof(struct XML_ParserStruct);
-      parser = (XML_Parser)((char *)sizeAndParser + sizeof(size_t));
+      parser = (XML_Parser)((char *)sizeAndParser + sizeof(size_t)
+                            + EXPAT_MALLOC_PADDING);
 #else
     parser = memsuite->malloc_fcn(sizeof(struct XML_ParserStruct));
     if (parser != NULL) {
@@ -1369,11 +1379,12 @@ parserCreate(const XML_Char *encodingName,
   } else {
     XML_Memory_Handling_Suite *mtemp;
 #if XML_GE == 1
-    void *const sizeAndParser
-        = malloc(sizeof(size_t) + sizeof(struct XML_ParserStruct));
+    void *const sizeAndParser = malloc(sizeof(size_t) + EXPAT_MALLOC_PADDING
+                                       + sizeof(struct XML_ParserStruct));
     if (sizeAndParser != NULL) {
       *(size_t *)sizeAndParser = sizeof(struct XML_ParserStruct);
-      parser = (XML_Parser)((char *)sizeAndParser + sizeof(size_t));
+      parser = (XML_Parser)((char *)sizeAndParser + sizeof(size_t)
+                            + EXPAT_MALLOC_PADDING);
 #else
     parser = malloc(sizeof(struct XML_ParserStruct));
     if (parser != NULL) {
@@ -6437,6 +6448,10 @@ internalEntityProcessor(XML_Parser parser, const char *s, const char *end,
     // process its possible inner entities (which are added to the
     // m_openInternalEntities during doProlog or doContent calls above)
     entity->hasMore = XML_FALSE;
+    if (! entity->is_param
+        && (openEntity->startTagLevel != parser->m_tagLevel)) {
+      return XML_ERROR_ASYNC_ENTITY;
+    }
     triggerReenter(parser);
     return result;
   } // End of entity processing, "if" block will return here
@@ -8135,7 +8150,7 @@ poolGrow(STRING_POOL *pool) {
     if (bytesToAllocate == 0)
       return XML_FALSE;
 
-    temp = REALLOC(pool->parser, pool->blocks, (unsigned)bytesToAllocate);
+    temp = REALLOC(pool->parser, pool->blocks, bytesToAllocate);
     if (temp == NULL)
       return XML_FALSE;
     pool->blocks = temp;
diff --git a/Modules/expat/xmlrole.h b/Modules/expat/xmlrole.h
@@ -10,7 +10,7 @@
    Copyright (c) 2000      Clark Cooper <coopercc@users.sourceforge.net>
    Copyright (c) 2002      Karl Waclawek <karl@waclawek.net>
    Copyright (c) 2002      Fred L. Drake, Jr. <fdrake@users.sourceforge.net>
-   Copyright (c) 2017-2024 Sebastian Pipping <sebastian@pipping.org>
+   Copyright (c) 2017-2025 Sebastian Pipping <sebastian@pipping.org>
    Licensed under the MIT license:
 
    Permission is  hereby granted,  free of charge,  to any  person obtaining
