I tried adding back Xalan and applying the patch mentioned.
I double checked the version of Xalan by default in the JDK (shaded under
com.sun.org.apache ...).

It's a Xalan 2.7.2 so that's the one I used as a starting point.

--
Jean-Louis Monteiro
http://twitter.com/jlouismonteiro
http://www.tomitribe.com


On Tue, Dec 8, 2020 at 3:09 PM Jonathan Gallimore <
jonathan.gallim...@gmail.com> wrote:

> We're looking to pass the TCK for both EE8 and EE9. This change would be
> needed on both counts.
>
> Both TomEE 8 and 9 use the exact same codebase, and one is simply a
> translation of the other into the jakarta. namespace.
>
> I understand your view, and also the issue you experienced with the
> Japanese character set. From your point of view, what would you like from
> the project to assure you that the problems you've previously seen don't
> appear again (or cause other issues)? We can definitely expand on tests in
> the project to increase levels of confidence and prevent regressions, I
> just think we need to agree on what those tests are.
>
> If I find another approach that can work, I'll present it here.
>
> Jon
>
> On Tue, Dec 8, 2020 at 1:46 PM Alex The Rocker <alex.m3...@gmail.com>
> wrote:
>
> > In that case, could xalan-reintroduction be limited to TomEE 9.x ?
> > What's the urge with TomEE 8.x vs. impacts risks ?
> >
> > Le mar. 8 déc. 2020 à 14:39, Jonathan Gallimore
> > <jonathan.gallim...@gmail.com> a écrit :
> > >
> > > We fixed a specific issue by removing Xalan, but without realizing the
> > > impact on JSTL. We'll need to fix this one way or another. There's a
> > couple
> > > of options:
> > >
> > > 1. Re-introduce Xalan and patch it (which seems to be the common
> > > approach across the app servers)
> > > 2. Implement what we need with JAXP.
> > >
> > > At the moment, I'm more in favour of (1) for 2 reasons: its more of a
> > known
> > > quantity (we've used it before) than implementing our thing, and
> > secondly,
> > > it keeps us in line with Tomcat. Our mantra has always been "Be
> Tomcat" -
> > > if we went the JAXP route we're consciously doing something a little
> > > different, and going against the idea of "Be Tomcat".
> > >
> > > Jon
> > >
> > > On Tue, Dec 8, 2020 at 12:50 PM Alex The Rocker <alex.m3...@gmail.com>
> > > wrote:
> > >
> > > > ForEachTag's class comment says "Implementation of <x:forEach>; tag
> > > > using low-level Xalan API."
> > > > Isn't there a way to emulate that with simple JAXP instead of
> > > > re-introducing the huge & unmaintained dependency on Xalan ?
> > > > As far as I remember, not only xalan.jar, but also serializer.jar had
> > > > to be removed (maybe in our apps, I don't remember), and that was a
> > > > bit chaotic. I'm scared at the idea of re-doing this error-prone
> > > > process in the reverse order, and who know how many unsolved bugs
> > > > Xalan fixed in vanilla JAXP will re-introduced on TomEE?
> > > >
> > > > Alexandre
> > > >
> > > > Le mar. 8 déc. 2020 à 13:39, Jean-Louis Monteiro
> > > > <jlmonte...@tomitribe.com> a écrit :
> > > > >
> > > > > Quick update on this one ...
> > > > >
> > > > > Tomcat Taglib has a direct dependency on XParthContext
> > > > >
> > > >
> >
> https://github.com/apache/tomcat-taglibs-standard/blob/master/impl/src/main/java/org/apache/taglibs/standard/tag/common/xml/ForEachTag.java
> > > > >
> > > > > This is available in Xalan
> > > > > So at this point, we have an issue in TomEE with xml:transform not
> > > > working.
> > > > > TCK or not, does not matter, it's not working.
> > > > >
> > > > > Long term would probably be to get Tomcat Taglib implementation to
> > use
> > > > JAXP
> > > > > APIs instead of straight Xalan.
> > > > > Short term would be to patch Xalan ourselves.
> > > > >
> > > > > Thoughts?
> > > > >
> > > > >
> > > > > --
> > > > > Jean-Louis Monteiro
> > > > > http://twitter.com/jlouismonteiro
> > > > > http://www.tomitribe.com
> > > > >
> > > > >
> > > > > On Tue, Dec 8, 2020 at 11:31 AM Jean-Louis Monteiro <
> > > > > jlmonte...@tomitribe.com> wrote:
> > > > >
> > > > > > Thanks.
> > > > > > Alexandre also replied on the ticket so I'll revert and see if I
> > can
> > > > find
> > > > > > a better solution.
> > > > > >
> > > > > > Thanks guys
> > > > > > --
> > > > > > Jean-Louis Monteiro
> > > > > > http://twitter.com/jlouismonteiro
> > > > > > http://www.tomitribe.com
> > > > > >
> > > > > >
> > > > > > On Tue, Dec 8, 2020 at 11:21 AM Jonathan Gallimore <
> > > > > > jonathan.gallim...@gmail.com> wrote:
> > > > > >
> > > > > >> Yes - there's a bug with parsing Japanese characters with Xalan.
> > It
> > > > also
> > > > > >> looks like its not maintained any more.
> > > > > >>
> > > > > >> I was under the impression that the Transformer functionality
> was
> > > > part of
> > > > > >> the JDK, so I'm curious as to what is missing. I'll have a look
> at
> > > > that
> > > > > >> failing test. Looks like we're missing org.apache.xpath - looks
> > like
> > > > this
> > > > > >> is a little messy.
> > > > > >>
> > > > > >> Jon
> > > > > >>
> > > > > >> On Tue, Dec 8, 2020 at 9:58 AM Jean-Louis Monteiro <
> > > > > >> jlmonte...@tomitribe.com>
> > > > > >> wrote:
> > > > > >>
> > > > > >> > Hi,
> > > > > >> >
> > > > > >> > Been looking at the JSTL failures
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > >
> >
> http://ec2-54-173-218-40.compute-1.amazonaws.com:17171/tests?path=com.sun.ts.tests.jstl&build=1607347669779&status=FAILED
> > > > > >> >
> > > > > >> > I realized that Xalan is missing from our distribution.
> > > > > >> > Looks like it's been removed about a year ago.
> > > > > >> >
> > > > > >> >
> > > > > >> >
> > > > > >>
> > > >
> >
> https://github.com/apache/tomee/commit/5f1b57af1b4f11638c8c9540fcc83feb618980a4
> > > > > >> >
> > > > > >> > Does anyone have any clue on why?
> > > > > >> > Obviously we won't pass the transform TCK without an XSLT
> > Processor
> > > > like
> > > > > >> > Xalan. Adding it back solved most of the issues.
> > > > > >> >
> > > > > >> > --
> > > > > >> > Jean-Louis Monteiro
> > > > > >> > http://twitter.com/jlouismonteiro
> > > > > >> > http://www.tomitribe.com
> > > > > >> >
> > > > > >>
> > > > > >
> > > >
> >
>

Reply via email to