Mono Test Suite Results

[Eoin Coffey]

As promised here are the test suite results. The following five are failing:

* API.Tests.EncodingTests.Japanese_ShiftJIS:

Looks like mono doesn't support the encoding 'Shift_JIS'. Not sure
what the resolution for this will be.

* API.Tests.IgnoreTests.TestIgnore:

Assert.AreEqual(true, rules.IgnoreDir("project/", "project/src/bin"));
is failing.

Also I suggest that we aim to have tests that contain only one Assert.
Helps focus on what you're really attempting to do. So for instance
in IgnoreTests we'd have a IgnoresGlob, IgnoresFile,
IgnoresFileNegation, IgnoresDirectory, IgnoresDirectoryNegation.

* API.Tests.RepositoryTests.FindRepository:

Assert.AreEqual(null, Repository.FindRepository("Q://a path that
surely does not exist " + Path.GetRandomFileName())); fails.

* Core.Tests.RepositoryCacheTest.testFileKeyOpenNew:

Assert.IsFalse(gitdir.Exists) fails.

* Core.Tests.RevWalk.RevCommitParseTest.testParse_explicit_bad_encoded:

Looks like another System.Text.Encoding mono "bug"

I've also attached the xml bomb nunit produces.

-Eoin

[Emeric Fermas]

Eoin,

Thanks a lot for having run them.

>* API.Tests.EncodingTests.Japanese_ShiftJIS:

This failure surprises me a bit. i do not remember having it failed before. What distro are you running ? what is your Mono version ?

@nestalk,

Do you also encounter this problem ?

>* API.Tests.IgnoreTests.TestIgnore:

This is known and related to a more broader issue which deal with complete .gitignore files and configuration settings handling (http://code.google.com/p/gitsharp/issues/detail?id=46)

>* API.Tests.RepositoryTests.FindRepository:

Hum... This one has clearly been written on a Windows OS ;-)

However, even if the path is nonsense on a Mono env, the FindRepository() implementation should not be able to send back the current GitSharp directory by walking up the file hierarchy.

I wonder how this piece of code " var git = Path.Combine(directory, Constants.DOT_GIT);" behaves given the Windows-like directory on Mono. 

Would you agree with taking a look at this ?

>* Core.Tests.RepositoryCacheTest.testFileKeyOpenNew:

This should be fixed and pushed up by now.

>*Core.Tests.RevWalk.RevCommitParseTest.testParse_explicit_bad_encoded:

Well there is a known problem with this one on Mono. In fact, one of the asserts makes the test being ignore when reaching a certain line of code which is desperately failing on Mono.

Nonetheless, my bet is the failure is being triggered by the parseCanonical() call. 

It looks that your Mono box does not look like working swiftly with exotic Encodings (In this case the EUC_JP....). Any idea ? 

Once again, thx for the help!

Em.

--
For infos about GitSharp visit http://www.eqqon.com/index.php/GitSharp
To join this group visit http://groups.google.com/group/gitsharp

[Eoin Coffey]

On Sat, Jan 23, 2010 at 8:14 AM, Emeric Fermas <emeric...@gmail.com> wrote:
> Eoin,
> Thanks a lot for having run them.

Yeah no problem

>>* API.Tests.EncodingTests.Japanese_ShiftJIS:
> This failure surprises me a bit. i do not remember having it failed before.
> What distro are you running ? what is your Mono version ?

Ubuntu 9.10 - kernel version: 2.6.31-17-generic-pae

mono --version:

Mono JIT compiler version 2.4.2.3 (Debian 2.4.2.3+dfsg-2)
Copyright (C) 2002-2008 Novell, Inc and Contributors. www.mono-project.com
TLS: __thread
GC: Included Boehm (with typed GC)
SIGSEGV: altstack
Notifications: epoll
Architecture: x86
Disabled: none

> @nestalk,
> Do you also encounter this problem ?
>

Could be I do not have some "esoteric" (well what I would consider
esoteric :-P) locales installed. A little googling for 'mono EUC-JP'
reveals this bug created by Neil in October:
https://bugzilla.novell.com/show_bug.cgi?id=547902 which only seems
vaguely related.

@nestalk, I am unable to reproduce that ticket. Instead of seeing the
string printed like you reported in actual result, I get the same
ArgumentException 'EUC-JP' not supported from the unit test failing.
Is there something besides the core mono libs you need to install to
get that encoding?

>>* API.Tests.IgnoreTests.TestIgnore:
> This is known and related to a more broader issue which deal with complete
> .gitignore files and configuration settings handling
> (http://code.google.com/p/gitsharp/issues/detail?id=46)
>>* API.Tests.RepositoryTests.FindRepository:
> Hum... This one has clearly been written on a Windows OS ;-)
> However, even if the path is nonsense on a Mono env, the FindRepository()
> implementation should not be able to send back the current GitSharp
> directory by walking up the file hierarchy.
> I wonder how this piece of code " var git = Path.Combine(directory,
> Constants.DOT_GIT);" behaves given the Windows-like directory on Mono.
> Would you agree with taking a look at this ?

Yeah I can dive into this :-)

>>* Core.Tests.RepositoryCacheTest.testFileKeyOpenNew:
> This should be fixed and pushed up by now.
>>*Core.Tests.RevWalk.RevCommitParseTest.testParse_explicit_bad_encoded:
> Well there is a known problem with this one on Mono. In fact, one of the
> asserts makes the test being ignore when reaching a certain line of code
> which is desperately failing on Mono.
> Nonetheless, my bet is the failure is being triggered by the
> parseCanonical() call.
> It looks that your Mono box does not look like working swiftly with exotic
> Encodings (In this case the EUC_JP....). Any idea ?

Can only assume this is related to the other encoding woes from the
Shift_JIS test.

[Eoin Coffey]

Was missing a mono specific package. So both Shift_JIS and
test_parse_explicit_bad_encoded are passing now.

[Emeric Fermas]

>Was missing a mono specific package.  So both Shift_JIS and
>test_parse_explicit_bad_encoded are passing now.

That's good news :-)


Em.