I think I do remember us using HTML Tidy as an HTML parser for the test cases, so that seems to make sense. Running the failing test case under a debugger ought to allow confirming that.
Re suppressing the message: one _could_ run tidy with its stdout/stderr captured and expected messages filtered out. Some newer test frameworks have that concept built into them, but it shouldn't be hard to implement if ours doesn't. -- /_ Joe Kesselman (he/him/his) -/ _) My Alexa skill for New Music/New Sounds fans: / https://www.amazon.com/dp/B09WJ3H657/ () Plaintext Ribbon Campaign /\ Stamp out HTML mail! ________________________________ From: Mukul Gandhi <muk...@apache.org> Sent: Monday, February 13, 2023 12:26:18 PM To: dev@xalan.apache.org <dev@xalan.apache.org> Subject: Re: [request to review, and vote] XalanJ 2.7.3 release candidate RC9 Hi Elliotte, On Mon, Feb 13, 2023 at 5:39 PM Elliotte Rusty Harold <elh...@ibiblio.org> wrote: > > This is a long shot but any chance HTML Tidy is in play here? I think > that has the string you're looking for. You could possibly be right, on these points. I think, HTML Tidy is used by XalanJ tests codebase, to construct a org.w3c.dom.Document object from an HTML document (produced by certain XalanJ tests). It seems that, certain XalanJ tests uses these org.w3c.dom.Document objects to do XML document tree comparisons to compare expected XSLT transform results with actual XSLT transform results. I've a feeling that, it may be hard to fix XalanJ tests error messages like "[xalantest] line 1 column 157 - Error: <append> is not recognized!", by trying to solve this with HTML Tidy. According to HTML Tidy's web site (https://github.com/jtidy/jtidy), which says "Checkout v.Nu validator for a possible modern replacement" (whose link is https://about.validator.nu/htmlparser/) at the bottom of its page. I think, we could explore using v.Nu validator to fix these issues, that we've been discussing within this thread. I'm just curious, should we try, doing these improvements on XalanJ tests implementation (for e.g, trying to use v.Nu validator with XalanJ tests) and then release XalanJ 2.7.3 with these fixes, or should we explore these fixes after we've released XalanJ 2.7.3? Having explored the XalanJ tests codebase up to now, I don't think, integrating v.Nu validator with XalanJ tests, should be a big hassle. But that, should take few more days, to complete doing these fixes. @ Gary and others, do you as well, have any thoughts on these points? -- Regards, Mukul Gandhi --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@xalan.apache.org For additional commands, e-mail: dev-h...@xalan.apache.org
