From 38f04779f7afd758db6210123ec0b64c489595c5 Mon Sep 17 00:00:00 2001 From: Nick Wellnhofer Date: Mon, 22 Aug 2022 13:33:35 +0200 Subject: [PATCH] Fix HTML parser with threads and --without-legacy If the legacy functions are disabled, the default "V1" HTML SAX handler isn't initialized in threads other than the main thread. htmlInitParserCtxt would later use the empty V1 SAX handler, resulting in NULL documents. Change htmlInitParserCtxt to initialize the HTML SAX handler by calling xmlSAX2InitHtmlDefaultSAXHandler. This removes the ability to change the default handler but is more in line with the XML parser which initializes the SAX handler by calling xmlSAXVersion, ignoring the V1 default handler. Fixes #399. Reference:https://github.com/GNOME/libxml2/commit/38f04779f7afd758db6210123ec0b64c489595c5 Conflict:NA --- HTMLparser.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) diff --git a/HTMLparser.c b/HTMLparser.c index e95d86b..98d73f3 100644 --- a/HTMLparser.c +++ b/HTMLparser.c @@ -5039,8 +5039,7 @@ htmlInitParserCtxt(htmlParserCtxtPtr ctxt) htmlErrMemory(NULL, "htmlInitParserCtxt: out of memory\n"); return(-1); } - else - memset(sax, 0, sizeof(htmlSAXHandler)); + memset(sax, 0, sizeof(htmlSAXHandler)); /* Allocate the Input stack */ ctxt->inputTab = (htmlParserInputPtr *) @@ -5099,11 +5098,9 @@ htmlInitParserCtxt(htmlParserCtxtPtr ctxt) ctxt->nodeInfoNr = 0; ctxt->nodeInfoMax = 0; - if (sax == NULL) ctxt->sax = (xmlSAXHandlerPtr) &htmlDefaultSAXHandler; - else { - ctxt->sax = sax; - memcpy(sax, &htmlDefaultSAXHandler, sizeof(xmlSAXHandlerV1)); - } + ctxt->sax = sax; + xmlSAX2InitHtmlDefaultSAXHandler(sax); + ctxt->userData = ctxt; ctxt->myDoc = NULL; ctxt->wellFormed = 1; -- 2.27.0