Only register error handling when observable

nielsdos · nielsdos · commit b95597381880 · 2024-03-17T18:24:40.000+01:00
Closes GH-13702.
diff --git a/UPGRADING.INTERNALS b/UPGRADING.INTERNALS
@@ -185,6 +185,7 @@ PHP 8.4 INTERNALS UPGRADE NOTES
    - Removed the "properties" HashTable field from php_libxml_node_object.
    - Added a way to attached private data to a php_libxml_ref_obj.
    - Added a way to fix a class type onto php_libxml_ref_obj.
+   - Added php_libxml_uses_internal_errors().
 
  e. ext/date
    - Added the php_format_date_ex() API to format instances of php_date_obj.
diff --git a/ext/dom/html_document.c b/ext/dom/html_document.c
@@ -765,6 +765,16 @@ PHP_METHOD(DOM_HTMLDocument, createEmpty)
 	RETURN_THROWS();
 }
 
+/* Only bother to register error handling when the error reports can become observable. */
+static bool dom_should_register_error_handlers(zend_long options)
+{
+	if (options & XML_PARSE_NOERROR) {
+		return false;
+	}
+
+	return php_libxml_uses_internal_errors() || ((EG(error_reporting) | EG(user_error_handler_error_reporting)) & E_WARNING);
+}
+
 PHP_METHOD(DOM_HTMLDocument, createFromString)
 {
 	const char *source, *override_encoding = NULL;
@@ -793,7 +803,7 @@ PHP_METHOD(DOM_HTMLDocument, createFromString)
 	dom_reset_line_column_cache(&application_data.cache_tokenizer);
 	lexbor_libxml2_bridge_parse_context ctx;
 	lexbor_libxml2_bridge_parse_context_init(&ctx);
-	if (!(options & XML_PARSE_NOERROR)) {
+	if (dom_should_register_error_handlers(options)) {
 		lexbor_libxml2_bridge_parse_set_error_callbacks(
 			&ctx,
 			dom_lexbor_libxml2_bridge_tokenizer_error_reporter,
@@ -952,7 +962,7 @@ PHP_METHOD(DOM_HTMLDocument, createFromFile)
 	dom_reset_line_column_cache(&application_data.cache_tokenizer);
 	lexbor_libxml2_bridge_parse_context ctx;
 	lexbor_libxml2_bridge_parse_context_init(&ctx);
-	if (!(options & XML_PARSE_NOERROR)) {
+	if (dom_should_register_error_handlers(options)) {
 		lexbor_libxml2_bridge_parse_set_error_callbacks(
 			&ctx,
 			dom_lexbor_libxml2_bridge_tokenizer_error_reporter,
diff --git a/ext/dom/tests/modern/html/parser/user_error_handler.phpt b/ext/dom/tests/modern/html/parser/user_error_handler.phpt
@@ -0,0 +1,19 @@
+--TEST--
+Parse HTML document with user error handler and error_reporting(0)
+--EXTENSIONS--
+dom
+--INI--
+error_reporting=0
+--FILE--
+<?php
+
+set_error_handler(function ($errno, $errstr, $errfile, $errline) {
+    var_dump($errno, $errstr);
+}, E_WARNING);
+
+DOM\HTMLDocument::createFromString('<html></html>');
+
+?>
+--EXPECT--
+int(2)
+string(113) "DOM\HTMLDocument::createFromString(): tree error unexpected-token-in-initial-mode in Entity, line: 1, column: 2-5"
diff --git a/ext/libxml/libxml.c b/ext/libxml/libxml.c
@@ -1055,23 +1055,22 @@ PHP_FUNCTION(libxml_set_streams_context)
 }
 /* }}} */
 
+PHP_LIBXML_API bool php_libxml_uses_internal_errors(void)
+{
+	return xmlStructuredError == php_libxml_structured_error_handler;
+}
+
 /* {{{ Disable libxml errors and allow user to fetch error information as needed */
 PHP_FUNCTION(libxml_use_internal_errors)
 {
-	xmlStructuredErrorFunc current_handler;
-	bool use_errors, use_errors_is_null = 1, retval;
+	bool use_errors, use_errors_is_null = true;
 
 	ZEND_PARSE_PARAMETERS_START(0, 1)
 		Z_PARAM_OPTIONAL
 		Z_PARAM_BOOL_OR_NULL(use_errors, use_errors_is_null)
 	ZEND_PARSE_PARAMETERS_END();
 
-	current_handler = xmlStructuredError;
-	if (current_handler && current_handler == php_libxml_structured_error_handler) {
-		retval = 1;
-	} else {
-		retval = 0;
-	}
+	bool retval = php_libxml_uses_internal_errors();
 
 	if (use_errors_is_null) {
 		RETURN_BOOL(retval);
diff --git a/ext/libxml/php_libxml.h b/ext/libxml/php_libxml.h
@@ -163,6 +163,7 @@ PHP_LIBXML_API void php_libxml_issue_error(int level, const char *msg);
 PHP_LIBXML_API bool php_libxml_disable_entity_loader(bool disable);
 PHP_LIBXML_API void php_libxml_set_old_ns(xmlDocPtr doc, xmlNsPtr ns);
 PHP_LIBXML_API php_stream_context *php_libxml_get_stream_context(void);
+PHP_LIBXML_API bool php_libxml_uses_internal_errors(void);
 
 PHP_LIBXML_API zend_string *php_libxml_sniff_charset_from_string(const char *start, const char *end);
 PHP_LIBXML_API zend_string *php_libxml_sniff_charset_from_stream(const php_stream *s);
