Update libxml2 to v2.12.5

.travis.yml

@@ -109,7 +109,7 @@ before_install:
     esac
   - curl https://bitbucket.org/Swyter/bitbucket-curl-upload-to-repo-downloads/raw/default/upload-to-bitbucket.sh -O -J -L
   - chmod +x ./upload-to-bitbucket.sh
-  - git clone --branch v2.12.0 --depth 1 https://gitlab.gnome.org/GNOME/libxml2
+  - git clone --branch v2.12.5 --depth 1 https://gitlab.gnome.org/GNOME/libxml2
   - mkdir -p libxml2/$PLATFORM
 
 before_cache:

ExternData/Resources/C-Sources/libxml2/HTMLparser.c

@@ -4851,6 +4851,14 @@ htmlParseDocument(htmlParserCtxtPtr ctxt) {
 
     xmlDetectEncoding(ctxt);
 
+    /*
+     * This is wrong but matches long-standing behavior. In most cases,
+     * a document starting with an XML declaration will specify UTF-8.
+     */
+    if (((ctxt->input->flags & XML_INPUT_HAS_ENCODING) == 0) &&
+        (xmlStrncmp(ctxt->input->cur, BAD_CAST "<?xm", 4) == 0))
+        xmlSwitchEncoding(ctxt, XML_CHAR_ENCODING_UTF8);
+
     /*
      * Wipe out everything which is before the first '<'
      */
@@ -5408,6 +5416,16 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
 		 */
 	        goto done;
             case XML_PARSER_START:
+                /*
+                 * This is wrong but matches long-standing behavior. In most
+                 * cases, a document starting with an XML declaration will
+                 * specify UTF-8.
+                 */
+                if (((ctxt->input->flags & XML_INPUT_HAS_ENCODING) == 0) &&
+                    (xmlStrncmp(ctxt->input->cur, BAD_CAST "<?xm", 4) == 0)) {
+                    xmlSwitchEncoding(ctxt, XML_CHAR_ENCODING_UTF8);
+                }
+
 	        /*
 		 * Very first chars read from the document flow.
 		 */
@@ -6823,7 +6841,8 @@ htmlCtxtReadMemory(htmlParserCtxtPtr ctxt, const char *buffer, int size,
 
     htmlCtxtReset(ctxt);
 
-    input = xmlParserInputBufferCreateMem(buffer, size, XML_CHAR_ENCODING_NONE);
+    input = xmlParserInputBufferCreateStatic(buffer, size,
+                                             XML_CHAR_ENCODING_NONE);
     if (input == NULL) {
 	return(NULL);
     }

ExternData/Resources/C-Sources/libxml2/SAX2.c

@@ -1897,8 +1897,17 @@ xmlSAX2AttributeNs(xmlParserCtxtPtr ctxt,
     memset(ret, 0, sizeof(xmlAttr));
     ret->type = XML_ATTRIBUTE_NODE;
 
+    /*
+     * xmlParseBalancedChunkMemoryRecover had a bug that could result in
+     * a mismatch between ctxt->node->doc and ctxt->myDoc. We use
+     * ctxt->node->doc here, but we should somehow make sure that the
+     * document pointers match.
+     */
+
+    /* assert(ctxt->node->doc == ctxt->myDoc); */
+
     ret->parent = ctxt->node;
-    ret->doc = ctxt->myDoc;
+    ret->doc = ctxt->node->doc;
     ret->ns = namespace;
 
     if (ctxt->dictNames)

ExternData/Resources/C-Sources/libxml2/dict.c

@@ -909,8 +909,8 @@ static xmlMutex xmlRngMutex;
 static unsigned globalRngState[2];
 
 #ifdef XML_THREAD_LOCAL
-XML_THREAD_LOCAL static int localRngInitialized = 0;
-XML_THREAD_LOCAL static unsigned localRngState[2];
+static XML_THREAD_LOCAL int localRngInitialized = 0;
+static XML_THREAD_LOCAL unsigned localRngState[2];
 #endif
 
 ATTRIBUTE_NO_SANITIZE_INTEGER

ExternData/Resources/C-Sources/libxml2/globals.c

@@ -82,7 +82,6 @@ XML_GLOBALS_ERROR
 XML_GLOBALS_HTML
 XML_GLOBALS_IO
 XML_GLOBALS_PARSER
-XML_GLOBALS_SAVE
 XML_GLOBALS_TREE
 #undef XML_OP
 };
@@ -100,8 +99,14 @@ static xmlMutex xmlThrDefMutex;
  * On Darwin, thread-local storage destructors seem to be run before
  * pthread thread-specific data destructors. This causes ASan to
  * report a use-after-free.
+ *
+ * On Windows, we can't use TLS in static builds. The RegisterWait
+ * callback would run after TLS was deallocated.
  */
-#if defined(XML_THREAD_LOCAL) && !defined(__APPLE__)
+#if defined(XML_THREAD_LOCAL) && \
+    !defined(__APPLE__) && \
+    (!defined(HAVE_WIN32_THREADS) || \
+     !defined(LIBXML_STATIC) || defined(LIBXML_STATIC_FOR_DLL))
 #define USE_TLS
 #endif
 
@@ -118,6 +123,13 @@ static XML_THREAD_LOCAL xmlGlobalState globalState;
     defined(__GLIBC__) && \
     __GLIBC__ * 100 + __GLIBC_MINOR__ < 234
 
+#pragma weak pthread_getspecific
+#pragma weak pthread_setspecific
+#pragma weak pthread_key_create
+#pragma weak pthread_key_delete
+#pragma weak pthread_equal
+#pragma weak pthread_self
+
 #define XML_PTHREAD_WEAK
 
 static int libxml_is_threaded = -1;
@@ -566,7 +578,13 @@ void xmlInitGlobalsInternal(void) {
             (pthread_getspecific != NULL) &&
             (pthread_setspecific != NULL) &&
             (pthread_key_create != NULL) &&
-            (pthread_key_delete != NULL);
+            (pthread_key_delete != NULL) &&
+            /*
+             * pthread_equal can be inline, resuting in -Waddress warnings.
+             * Let's assume it's available if all the other functions are.
+             */
+            /* (pthread_equal != NULL) && */
+            (pthread_self != NULL);
     if (libxml_is_threaded == 0)
         return;
 #endif /* XML_PTHREAD_WEAK */
@@ -889,7 +907,6 @@ XML_GLOBALS_ERROR
 XML_GLOBALS_HTML
 XML_GLOBALS_IO
 XML_GLOBALS_PARSER
-XML_GLOBALS_SAVE
 XML_GLOBALS_TREE
 #undef XML_OP
 

ExternData/Resources/C-Sources/libxml2/hash.c

@@ -913,15 +913,42 @@ xmlHashScan(xmlHashTablePtr hash, xmlHashScanner scan, void *data) {
 void
 xmlHashScanFull(xmlHashTablePtr hash, xmlHashScannerFull scan, void *data) {
     const xmlHashEntry *entry, *end;
+    xmlHashEntry old;
+    unsigned i;
 
     if ((hash == NULL) || (hash->size == 0) || (scan == NULL))
         return;
 
+    /*
+     * We must handle the case that a scanned entry is removed when executing
+     * the callback (xmlCleanSpecialAttr and possibly other places).
+     *
+     * Find the start of a probe sequence to avoid scanning entries twice if
+     * a deletion happens.
+     */
+    entry = hash->table;
     end = &hash->table[hash->size];
+    while (entry->hashValue != 0) {
+        if (++entry >= end)
+            entry = hash->table;
+    }
 
-    for (entry = hash->table; entry < end; entry++) {
-        if ((entry->hashValue != 0) && (entry->payload != NULL))
-            scan(entry->payload, data, entry->key, entry->key2, entry->key3);
+    for (i = 0; i < hash->size; i++) {
+        if ((entry->hashValue != 0) && (entry->payload != NULL)) {
+            /*
+             * Make sure to rescan after a possible deletion.
+             */
+            do {
+                old = *entry;
+                scan(entry->payload, data, entry->key, entry->key2, entry->key3);
+            } while ((entry->hashValue != 0) &&
+                     (entry->payload != NULL) &&
+                     ((entry->key != old.key) ||
+                      (entry->key2 != old.key2) ||
+                      (entry->key3 != old.key3)));
+        }
+        if (++entry >= end)
+            entry = hash->table;
     }
 }
 
@@ -966,22 +993,47 @@ xmlHashScanFull3(xmlHashTablePtr hash, const xmlChar *key,
                  const xmlChar *key2, const xmlChar *key3,
                  xmlHashScannerFull scan, void *data) {
     const xmlHashEntry *entry, *end;
+    xmlHashEntry old;
+    unsigned i;
 
     if ((hash == NULL) || (hash->size == 0) || (scan == NULL))
         return;
 
+    /*
+     * We must handle the case that a scanned entry is removed when executing
+     * the callback (xmlCleanSpecialAttr and possibly other places).
+     *
+     * Find the start of a probe sequence to avoid scanning entries twice if
+     * a deletion happens.
+     */
+    entry = hash->table;
     end = &hash->table[hash->size];
+    while (entry->hashValue != 0) {
+        if (++entry >= end)
+            entry = hash->table;
+    }
 
-    for (entry = hash->table; entry < end; entry++) {
-        if (entry->hashValue == 0)
-            continue;
-        if (((key == NULL) ||
-             (strcmp((const char *) key, (const char *) entry->key) == 0)) &&
-            ((key2 == NULL) || (xmlFastStrEqual(key2, entry->key2))) &&
-            ((key3 == NULL) || (xmlFastStrEqual(key3, entry->key3))) &&
-            (entry->payload != NULL)) {
-            scan(entry->payload, data, entry->key, entry->key2, entry->key3);
+    for (i = 0; i < hash->size; i++) {
+        if ((entry->hashValue != 0) && (entry->payload != NULL)) {
+            /*
+             * Make sure to rescan after a possible deletion.
+             */
+            do {
+                if (((key != NULL) && (strcmp((const char *) key,
+                                              (const char *) entry->key) != 0)) ||
+                    ((key2 != NULL) && (!xmlFastStrEqual(key2, entry->key2))) ||
+                    ((key3 != NULL) && (!xmlFastStrEqual(key3, entry->key3))))
+                    break;
+                old = *entry;
+                scan(entry->payload, data, entry->key, entry->key2, entry->key3);
+            } while ((entry->hashValue != 0) &&
+                     (entry->payload != NULL) &&
+                     ((entry->key != old.key) ||
+                      (entry->key2 != old.key2) ||
+                      (entry->key3 != old.key3)));
         }
+        if (++entry >= end)
+            entry = hash->table;
     }
 }
 

ExternData/Resources/C-Sources/libxml2/include/libxml/entities.h

@@ -12,7 +12,9 @@
 #define __XML_ENTITIES_H__
 
 #include <libxml/xmlversion.h>
+#define XML_TREE_INTERNALS
 #include <libxml/tree.h>
+#undef XML_TREE_INTERNALS
 
 #ifdef __cplusplus
 extern "C" {

ExternData/Resources/C-Sources/libxml2/include/libxml/globals.h

@@ -29,10 +29,6 @@ extern "C" {
 typedef struct _xmlGlobalState xmlGlobalState;
 typedef xmlGlobalState *xmlGlobalStatePtr;
 
-XML_DEPRECATED XMLPUBFUN void
-xmlInitGlobals(void);
-XML_DEPRECATED XMLPUBFUN void
-xmlCleanupGlobals(void);
 XML_DEPRECATED XMLPUBFUN void
 xmlInitializeGlobalState(xmlGlobalStatePtr gs);
 XML_DEPRECATED XMLPUBFUN

ExternData/Resources/C-Sources/libxml2/include/libxml/parser.h

@@ -11,7 +11,9 @@
 #define __XML_PARSER_H__
 
 #include <libxml/xmlversion.h>
+#define XML_TREE_INTERNALS
 #include <libxml/tree.h>
+#undef XML_TREE_INTERNALS
 #include <libxml/dict.h>
 #include <libxml/hash.h>
 #include <libxml/valid.h>
@@ -745,6 +747,19 @@ struct _xmlSAXHandler {
     setDocumentLocatorSAXFunc setDocumentLocator;
     startDocumentSAXFunc startDocument;
     endDocumentSAXFunc endDocument;
+    /*
+     * `startElement` and `endElement` are only used by the legacy SAX1
+     * interface and should not be used in new software. If you really
+     * have to enable SAX1, the preferred way is set the `initialized`
+     * member to 1 instead of XML_SAX2_MAGIC.
+     *
+     * For backward compatibility, it's also possible to set the
+     * `startElementNs` and `endElementNs` handlers to NULL.
+     *
+     * You can also set the XML_PARSE_SAX1 parser option, but versions
+     * older than 2.12.0 will probably crash if this option is provided
+     * together with custom SAX callbacks.
+     */
     startElementSAXFunc startElement;
     endElementSAXFunc endElement;
     referenceSAXFunc reference;
@@ -758,8 +773,14 @@ struct _xmlSAXHandler {
     getParameterEntitySAXFunc getParameterEntity;
     cdataBlockSAXFunc cdataBlock;
     externalSubsetSAXFunc externalSubset;
+    /*
+     * `initialized` should always be set to XML_SAX2_MAGIC to enable the
+     * modern SAX2 interface.
+     */
     unsigned int initialized;
-    /* The following fields are extensions available only on version 2 */
+    /*
+     * The following members are only used by the SAX2 interface.
+     */
     void *_private;
     startElementNsSAX2Func startElementNs;
     endElementNsSAX2Func endElementNs;
@@ -840,6 +861,15 @@ XMLPUBFUN const char *const *__xmlParserVersion(void);
   XML_OP(xmlPedanticParserDefaultValue, int, XML_DEPRECATED) \
   XML_OP(xmlSubstituteEntitiesDefaultValue, int, XML_DEPRECATED)
 
+#ifdef LIBXML_OUTPUT_ENABLED
+  #define XML_GLOBALS_PARSER_OUTPUT \
+    XML_OP(xmlIndentTreeOutput, int, XML_NO_ATTR) \
+    XML_OP(xmlTreeIndentString, const char *, XML_NO_ATTR) \
+    XML_OP(xmlSaveNoEmptyTags, int, XML_NO_ATTR)
+#else
+  #define XML_GLOBALS_PARSER_OUTPUT
+#endif
+
 #ifdef LIBXML_SAX1_ENABLED
   #define XML_GLOBALS_PARSER_SAX1 \
     XML_OP(xmlDefaultSAXHandler, xmlSAXHandlerV1, XML_DEPRECATED)
@@ -849,6 +879,7 @@ XMLPUBFUN const char *const *__xmlParserVersion(void);
 
 #define XML_GLOBALS_PARSER \
   XML_GLOBALS_PARSER_CORE \
+  XML_GLOBALS_PARSER_OUTPUT \
   XML_GLOBALS_PARSER_SAX1
 
 #define XML_OP XML_DECLARE_GLOBAL
@@ -872,6 +903,11 @@ XML_GLOBALS_PARSER
     XML_GLOBAL_MACRO(xmlPedanticParserDefaultValue)
   #define xmlSubstituteEntitiesDefaultValue \
     XML_GLOBAL_MACRO(xmlSubstituteEntitiesDefaultValue)
+  #ifdef LIBXML_OUTPUT_ENABLED
+    #define xmlIndentTreeOutput XML_GLOBAL_MACRO(xmlIndentTreeOutput)
+    #define xmlTreeIndentString XML_GLOBAL_MACRO(xmlTreeIndentString)
+    #define xmlSaveNoEmptyTags XML_GLOBAL_MACRO(xmlSaveNoEmptyTags)
+  #endif
 #endif
 /** DOC_ENABLE */
 
@@ -882,6 +918,12 @@ XMLPUBFUN void
 		xmlInitParser		(void);
 XMLPUBFUN void
 		xmlCleanupParser	(void);
+XML_DEPRECATED
+XMLPUBFUN void
+		xmlInitGlobals		(void);
+XML_DEPRECATED
+XMLPUBFUN void
+		xmlCleanupGlobals	(void);
 
 /*
  * Input functions

ExternData/Resources/C-Sources/libxml2/include/libxml/relaxng.h

@@ -13,6 +13,7 @@
 #include <libxml/xmlversion.h>
 #include <libxml/xmlerror.h>
 #include <libxml/xmlstring.h>
+#include <libxml/tree.h>
 
 #ifdef LIBXML_SCHEMAS_ENABLED
 

ExternData/Resources/C-Sources/libxml2/include/libxml/schemasInternals.h

@@ -21,6 +21,7 @@
 #include <libxml/xmlregexp.h>
 #include <libxml/hash.h>
 #include <libxml/dict.h>
+#include <libxml/tree.h>
 
 #ifdef __cplusplus
 extern "C" {

ExternData/Resources/C-Sources/libxml2/include/libxml/schematron.h

@@ -15,6 +15,7 @@
 
 #ifdef LIBXML_SCHEMATRON_ENABLED
 
+#include <libxml/xmlerror.h>
 #include <libxml/tree.h>
 
 #ifdef __cplusplus