Bug 118291 – dev-java/adaptx-0.9.13_p20041105-r1 fails with 1.5

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Miroslav Šulc 2006-01-08 06:20:35 UTC]

I had problem with dev-java/adaptx-0.9.13_p20041105-r1 compilation because of source version and TransformerFactoryImpl.java doesn't implement setFeature() method. I have created new ebuild and patches that changes source version to 1.4 and added blank implementation of setFeature() that throws exception.

Now it compiles on sun-jdk-1.5.0.06.

Java documentation says "All implementations are required to support the XMLConstants.FEATURE_SECURE_PROCESSING feature.". I didn't add the support because I don't know adaptx. What I did is I just solved the compilation problem.

I would appreciate if anyone who knows more about adaptx could review this patch and eventually correct it.

[Comment 1 Miroslav Šulc 2006-01-08 06:21:56 UTC]

Created attachment 76524 [details]
Ebuild using the patches

[Comment 2 Miroslav Šulc 2006-01-08 06:22:25 UTC]

Created attachment 76525 [details, diff]
Source version patch

[Comment 3 Miroslav Šulc 2006-01-08 06:23:06 UTC]

Created attachment 76526 [details, diff]
Adds empty implementation of setFeature()

[Comment 4 Josh Nichols (RETIRED) 2006-01-08 10:52:24 UTC]

(In reply to comment #0)
> I had problem with dev-java/adaptx-0.9.13_p20041105-r1 compilation because of
> source version and TransformerFactoryImpl.java doesn't implement setFeature()
> method. I have created new ebuild and patches that changes source version to
> 1.4 and added blank implementation of setFeature() that throws exception.
> 
> Now it compiles on sun-jdk-1.5.0.06.
> 
> Java documentation says "All implementations are required to support the
> XMLConstants.FEATURE_SECURE_PROCESSING feature.". I didn't add the support
> because I don't know adaptx. What I did is I just solved the compilation
> problem.
> 
> I would appreciate if anyone who knows more about adaptx could review this
> patch and eventually correct it.
> 

I think that taking this approach would be a bad idea adding empty implementations of functions needed for an inerface. A better one would be to take this up with upstream. The snapshot is pretty old, so it could well be that they've released a version which supports 1.5.

[Comment 5 Miroslav Šulc 2006-01-08 11:14:46 UTC]

I thought so but I couldn't find homepage of the project. The one that is mentioned in the ebuild didn't work and Google didn't find anything useful either.

[Comment 6 Petteri Räty (RETIRED) 2006-01-08 11:18:43 UTC]

The sources are here:
http://cvs.castor.codehaus.org/viewrep/~br=MAIN/castor/adaptx/adaptx/src

[Comment 7 Josh Nichols (RETIRED) 2006-01-12 16:07:26 UTC]

Removing [SOLVED] from summary, because the current solution is less than ideal.

[Comment 8 Josh Nichols (RETIRED) 2006-01-15 13:15:12 UTC]

I gave the latest code in CVS a whirl using 1.5, and it also suffers from the same problem.

[Comment 9 Josh Nichols (RETIRED) 2006-01-21 21:32:50 UTC]

Right now, I'm thinking that it would be better to throw UnsupportedOperationException in setFeature would be the best bet, until (and if...) upstream can fix it.

[Comment 10 Alon Bar-Lev (RETIRED) 2006-07-04 11:29:08 UTC]

Created attachment 90876 [details, diff]
adaptx-0.9.13_p20041105-jdk15.patch

No need for source.
A complete patch.

[Comment 11 Vlastimil Babka (Caster) (RETIRED) 2007-06-01 18:33:45 UTC]

I think we should (and in all similar cases) be throwing an AbstractMethodError. Because that's what you get if you compile it with 1.4 and then try to invoke the method added by 1.5 interface. I've tried it.

[Comment 12 Serkan Kaba (RETIRED) 2009-07-01 16:54:46 UTC]

Package removed from tree.