Great, Andreas! One of the many things I want to do, in my copious spare time, is to look at what changes Sun made to their copy of Xalan and try to make sure we have addressed the same points. I am really not delighted by the fact that they are shipping the code, and patching the code, without passing the patches back to Apache.
-- /_ Joe Kesselman (he/him/his) -/ _) My Alexa skill for New Music/New Sounds fans: / https://www.amazon.com/dp/B09WJ3H657/ Caveat: Opinionated old geezer with overcompensated writer's block. May be redundant, verbose, prolix, sesquipedalian, didactic, officious, or redundant. ________________________________ From: Andreas Martens <amart...@uk.ibm.com> Sent: Tuesday, April 1, 2025 10:50:28 AM To: dev@xalan.apache.org <dev@xalan.apache.org> Subject: Re: three PRs and counting Thanks both Greg and Joe, I’ve got xalan-test and all the important tests pass, I’ll get up to speed on writing new tests next! Since it’s in a separate repo I’ll create a separate PR for them when they’re ready :-) Cheers, Andreas From: Joseph Kesselman <kesh...@alum.mit.edu.INVALID> Date: Monday, 31 March 2025 at 14:20 To: dev@xalan.apache.org <dev@xalan.apache.org> Subject: [EXTERNAL] Re: three PRs and counting The test suite for Xalan still lives in a separate Apache project, xalan-test, since it had ambitions of being a test suite shared by other implementations of XSLT. If you downloaded one of the archive files (jar or zip), a copy of xalan-test The test suite for Xalan still lives in a separate Apache project, xalan-test, since it had ambitions of being a test suite shared by other implementations of XSLT. If you downloaded one of the archive files (jar or zip), a copy of xalan-test as of that release date should have been included; you can always fetch the current version from git. The test framework is still based on Ant rather than Maven. Note that some of the tests are known and accepted failures, where Xalan is known not to conform precisely to the formal XSLT Recommendation. The drivers should both report failure on those individual tests and success on the overall test run. Yes, any behavioral changes checked into the Xalan code base should be accompanied by new tests in xalan-tests to demonstrate the difference between old and new code and that the new function is working as expected, and of course should be shown not to break existing tests. -- /_ Joe Kesselman (he/him/his) -/ _) My Alexa skill for New Music/New Sounds fans: / https://www.amazon.com/dp/B09WJ3H657/ Caveat: Opinionated old geezer with overcompensated writer's block. May be redundant, verbose, prolix, sesquipedalian, didactic, officious, or redundant. ________________________________ From: Gary Gregory <garydgreg...@gmail.com> Sent: Monday, March 31, 2025 5:46:16 AM To: dev@xalan.apache.org <dev@xalan.apache.org> Subject: Re: three PRs and counting Hello Andreas, Thank you for your report and work. I am traveling for the next couple of days so I can't help much. You're in the right place to ask for help. Yes, adding tests to cover fixes and changes is important, otherwise reverting behavior is possible without realizing it. Gary On Mon, Mar 31, 2025, 10:02 Andreas Martens <amart...@uk.ibm.com<mailto:amart...@uk.ibm.com>> wrote: Hello all! Just been going through some of the problems our customers have hit whilst using our XSLT functionality that’s now underpin by XSLTC rather than IBMs XL-TXE. * https://github.com/apache/xalan-java/pull/189 fixes an unlimited growth of namespaces used within XSLTC, not a problem for short-lived programs, but we tend to run for months and years continuously processing! * https://github.com/apache/xalan-java/pull/192 fixes what I suspect was an accidental removal in XALANJ-2108 (in 2006!) of some code that ensured the variable vector was closed, which was discovered shortly after in 2008 in XALANJ-2430 (and more recently by our customer…) * https://github.com/apache/xalan-java/pull/193 adds some functionality to one of the shipped samples that I’ve found useful in debugging the two above, on the vein of if I found it useful, perhaps someone else did too! I ignored the one where the customer was referencing `msxsl:node-set($var)/Namespace` from `xmlns:msxsl="urn:schemas-microsoft-com:xslt"`, claiming that “it used to work” I presume there are some regression tests in there somewhere, so I’m going to see if I can find them and add my new fixes to them, so should hopefully come with another PR for that soon (or even better, extend existing PRs if I get there before reviews).. Anyway, how can I get these PRs reviewed and either approved or rejected? The last two are new, but the first one is from September, so I’d like to know what to do with it before I completely forget what I did :-) Cheers, Andreas -- Andreas Martens [signature_558150371] Senior Engineer App Connect Enterprise IBM Unless otherwise stated above: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: Building C, IBM Hursley Office, Hursley Park Road, Winchester, Hampshire SO21 2JN Unless otherwise stated above: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: Building C, IBM Hursley Office, Hursley Park Road, Winchester, Hampshire SO21 2JN
