https://bugzilla.wikimedia.org/show_bug.cgi?id=53410
Web browser: --- Bug ID: 53410 Summary: Don't modify the document after passing it to callbacks in mediawiki.HTML5TreeBuilder.node.js Product: Parsoid Version: unspecified Hardware: All OS: All Status: NEW Severity: normal Priority: Unprioritized Component: DOM Assignee: gwi...@wikimedia.org Reporter: gwi...@wikimedia.org CC: ssas...@wikimedia.org Classification: Unclassified Mobile Platform: --- When we process a document to HTML DOM, the document that is passed to callbacks is eventually modified in mediawiki.HTML5TreeBuilder.node.js when the pipeline is reused for another parse. This is unexpected behavior, so should really be fixed. The main reason for this is that there seems to be no efficient way to re-initialize an existing HTML5 parser with a new document. Creating a new HTML5 parser would work, but is (or at least was) quite expensive with a lot of work done in the constructor. So we should probably: 1) measure how expensive it currently is to create a new HTML5 parser from scratch, and if that is too slow, 2) create a way to efficiently reset the HTML5 parser with a new document without clobbering the old one. See also https://gerrit.wikimedia.org/r/#/c/81251/ for an earlier partial fix and bug 53407 for an example illustrating why modifying the document is a problem. -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list Wikibugs-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikibugs-l