[Mono-devel-list] Re: Fixes to XSLT

[Atsushi Eno]

Hi,

Andrew Skiba wrote:
> Atsushi Eno wrote:
> 
>> Okay, it also throws that exception in my box, but it is not reported
>> in your Class1.cs test ... ?
> 
> Class1.cs catches all the exceptions and prints the text of them into 
> the results file, if I understand correct your question.

Then, it should have been reported as differences, but I could not
find that specific test case.

>> So maybe there is no reference (original) "catalog.xml", right?
> 
> Sorry, I again do not understand what you just said. If you mean that 
> there is no original catalog.xml, it's a part of OASIS test suite. My 
> catalog-out.xml is produced from the original catalog by a simple XSL, 
> so it ended up in the same dir.

There is no "catalog.xml" in testsuite/TESTS/MSFT_Conformance_Tests
in that OASIS test archive, unless I am specially filtered ;-)

catalog.xml is only in testsuite/TESTS/Xalan_Conformance_Tests in
that archive, but I don't know if there used to be catalog.xml in
MS test dir in their old archives. In short, am afraid that you have
possiblly older archive.

>> Well, apparently Xalan test collection is less than MS tests, so
>> it is great that now we can easily test MS test collections with
>> your testcode ;-) I'll modify Class1.cs as runnable under 
>> "standalone_tests" (and file name as well ;-) and check into svn.
> 
> Cool! Than probably it's better to use my XSL on the up-to date OASIS 
> catalog.xml to make updated catalog-out.xml, see the attached file.

Again am mazed since there is no catalog.xml ... ;-)

>> BTW am on reviewing your patch and committed some of them. Now
>> the error (well, different result) cases are 438, from 557.
>> It is very nice :-) I'll continue my review and mail you later.
> 
> Yes, I hope that changes in this patch are very simple, so there will be 
> no problems with them. There is another patch, which I named 
> "notsogood.patch", there we will probably have problems.

cool. I have 4 remaining items to review in that patch. Will post
comments on patches that ended up remaining.

BTW it would be very nice if you split the patches per individual
problem you fixed. Also, "svn diff" format would be better.

> I'm still working on the enum issue, BTW, so you will hear from me today.

What I am really afraid is that the bug will result in incorrect
listing (as I wrote in MSDN feedback) of the enum type on reflection
based code editor like MonoDevelop. It is clear and present danger.

Atsushi Eno