gh-126623: Update libexpat to 2.6.4, make future updates easier (GH-126792)

Update libexpat to 2.6.4, make future updates easier.

4 files changed, 79 insertions, 11 deletions

diff --git a/Modules/expat/expat.h b/Modules/expat/expat.h
index d0d6015..523b37d 100644
--- a/Modules/expat/expat.h
+++ b/Modules/expat/expat.h
@@ -130,7 +130,9 @@ enum XML_Error {
   /* Added in 2.3.0. */
   XML_ERROR_NO_BUFFER,
   /* Added in 2.4.0. */
-  XML_ERROR_AMPLIFICATION_LIMIT_BREACH
+  XML_ERROR_AMPLIFICATION_LIMIT_BREACH,
+  /* Added in 2.6.4. */
+  XML_ERROR_NOT_STARTED,
 };
 
 enum XML_Content_Type {
@@ -1066,7 +1068,7 @@ XML_SetReparseDeferralEnabled(XML_Parser parser, XML_Bool enabled);
 */
 #define XML_MAJOR_VERSION 2
 #define XML_MINOR_VERSION 6
-#define XML_MICRO_VERSION 3
+#define XML_MICRO_VERSION 4
 
 #ifdef __cplusplus
 }
diff --git a/Modules/expat/expat_external.h b/Modules/expat/expat_external.h
index 12c560e..567872b 100644
--- a/Modules/expat/expat_external.h
+++ b/Modules/expat/expat_external.h
@@ -40,6 +40,10 @@
 #ifndef Expat_External_INCLUDED
 #define Expat_External_INCLUDED 1
 
+/* Namespace external symbols to allow multiple libexpat version to
+   co-exist. */
+#include "pyexpatns.h"
+
 /* External API definitions */
 
 /* Expat tries very hard to make the API boundary very specifically
@@ -64,11 +68,6 @@
    compiled with the cdecl calling convention as the default since
    system headers may assume the cdecl convention.
 */
-
-/* Namespace external symbols to allow multiple libexpat version to
-   co-exist. */
-#include "pyexpatns.h"
-
 #ifndef XMLCALL
 #  if defined(_MSC_VER)
 #    define XMLCALL __cdecl
diff --git a/Modules/expat/refresh.sh b/Modules/expat/refresh.sh
new file mode 100755
index 0000000..82a9dbc
--- /dev/null
+++ b/Modules/expat/refresh.sh
@@ -0,0 +1,57 @@
+#!/usr/bin/env bash
+#
+# Use this script to update libexpat
+
+set -e
+set -o pipefail
+
+if [[ "${BASH_VERSINFO[0]}" -lt 4 ]]; then
+  echo "A bash version >= 4 required. Got: $BASH_VERSION" >&2
+  exit 1
+fi
+
+# Update this when updating to a new version after verifying that the changes
+# the update brings in are good. These values are used for verifying the SBOM, too.
+expected_libexpat_tag="R_2_6_4"
+expected_libexpat_version="2.6.4"
+expected_libexpat_sha256="fd03b7172b3bd7427a3e7a812063f74754f24542429b634e0db6511b53fb2278"
+
+expat_dir="$(realpath "$(dirname -- "${BASH_SOURCE[0]}")")"
+cd ${expat_dir}
+
+# Step 1: download and copy files
+curl --location "https://github.com/libexpat/libexpat/releases/download/${expected_libexpat_tag}/expat-${expected_libexpat_version}.tar.gz" > libexpat.tar.gz
+echo "${expected_libexpat_sha256} libexpat.tar.gz" | sha256sum --check
+
+# Step 2: Pull files from the libexpat distribution
+declare -a lib_files
+lib_files=(
+  ascii.h
+  asciitab.h
+  expat.h
+  expat_external.h
+  iasciitab.h
+  internal.h
+  latin1tab.h
+  nametab.h
+  siphash.h
+  utf8tab.h
+  winconfig.h
+  xmlparse.c
+  xmlrole.c
+  xmlrole.h
+  xmltok.c
+  xmltok.h
+  xmltok_impl.c
+  xmltok_impl.h
+  xmltok_ns.c
+)
+for f in "${lib_files[@]}"; do
+  tar xzvf libexpat.tar.gz "expat-${expected_libexpat_version}/lib/${f}" --strip-components 2
+done
+rm libexpat.tar.gz
+
+# Step 3: Add the namespacing include to expat_external.h
+sed -i 's/#define Expat_External_INCLUDED 1/&\n\n\/* Namespace external symbols to allow multiple libexpat version to\n   co-exist. \*\/\n#include "pyexpatns.h"/' expat_external.h
+
+echo "Updated; verify all is okay using git diff and git status."
diff --git a/Modules/expat/xmlparse.c b/Modules/expat/xmlparse.c
index d9285b2..a4e091e 100644
--- a/Modules/expat/xmlparse.c
+++ b/Modules/expat/xmlparse.c
@@ -1,4 +1,4 @@
-/* ba4cdf9bdb534f355a9def4c9e25d20ee8e72f95b0a4d930be52e563f5080196 (2.6.3+)
+/* c5625880f4bf417c1463deee4eb92d86ff413f802048621c57e25fe483eb59e4 (2.6.4+)
                             __  __            _
                          ___\ \/ /_ __   __ _| |_
                         / _ \\  /| '_ \ / _` | __|
@@ -40,6 +40,7 @@
    Copyright (c) 2023      Owain Davies <owaind@bath.edu>
    Copyright (c) 2023-2024 Sony Corporation / Snild Dolkow <snild@sony.com>
    Copyright (c) 2024      Berkay Eren Ürün <berkay.ueruen@siemens.com>
+   Copyright (c) 2024      Hanno Böck <hanno@gentoo.org>
    Licensed under the MIT license:
 
    Permission is  hereby granted,  free of charge,  to any  person obtaining
@@ -2234,6 +2235,9 @@ XML_StopParser(XML_Parser parser, XML_Bool resumable) {
   if (parser == NULL)
     return XML_STATUS_ERROR;
   switch (parser->m_parsingStatus.parsing) {
+  case XML_INITIALIZED:
+    parser->m_errorCode = XML_ERROR_NOT_STARTED;
+    return XML_STATUS_ERROR;
   case XML_SUSPENDED:
     if (resumable) {
       parser->m_errorCode = XML_ERROR_SUSPENDED;
@@ -2244,7 +2248,7 @@ XML_StopParser(XML_Parser parser, XML_Bool resumable) {
   case XML_FINISHED:
     parser->m_errorCode = XML_ERROR_FINISHED;
     return XML_STATUS_ERROR;
-  default:
+  case XML_PARSING:
     if (resumable) {
 #ifdef XML_DTD
       if (parser->m_isParamEntity) {
@@ -2255,6 +2259,9 @@ XML_StopParser(XML_Parser parser, XML_Bool resumable) {
       parser->m_parsingStatus.parsing = XML_SUSPENDED;
     } else
       parser->m_parsingStatus.parsing = XML_FINISHED;
+    break;
+  default:
+    assert(0);
   }
   return XML_STATUS_OK;
 }
@@ -2519,6 +2526,9 @@ XML_ErrorString(enum XML_Error code) {
   case XML_ERROR_AMPLIFICATION_LIMIT_BREACH:
     return XML_L(
         "limit on input amplification factor (from DTD and entities) breached");
+  /* Added in 2.6.4. */
+  case XML_ERROR_NOT_STARTED:
+    return XML_L("parser not started");
   }
   return NULL;
 }
@@ -7856,7 +7866,7 @@ accountingReportDiff(XML_Parser rootParser,
   assert(! rootParser->m_parentParser);
 
   fprintf(stderr,
-          " (+" EXPAT_FMT_PTRDIFF_T("6") " bytes %s|%d, xmlparse.c:%d) %*s\"",
+          " (+" EXPAT_FMT_PTRDIFF_T("6") " bytes %s|%u, xmlparse.c:%d) %*s\"",
           bytesMore, (account == XML_ACCOUNT_DIRECT) ? "DIR" : "EXP",
           levelsAwayFromRootParser, source_line, 10, "");
 
@@ -7969,7 +7979,7 @@ entityTrackingReportStats(XML_Parser rootParser, ENTITY *entity,
 
   fprintf(
       stderr,
-      "expat: Entities(%p): Count %9d, depth %2d/%2d %*s%s%s; %s length %d (xmlparse.c:%d)\n",
+      "expat: Entities(%p): Count %9u, depth %2u/%2u %*s%s%s; %s length %d (xmlparse.c:%d)\n",
       (void *)rootParser, rootParser->m_entity_stats.countEverOpened,
       rootParser->m_entity_stats.currentDepth,
       rootParser->m_entity_stats.maximumDepthSeen,