[Mono-dev] Questions about Mono 'make check' [Fwd: About gentoo bug 257313]

[Leszek Ciesielski]

I am forwarding some questions I got from Gentoo community regarding
Mono build and test suite, I hope someone here can provide assistance.
Related Gentoo bugs are http://bugs.gentoo.org/show_bug.cgi?id=257313
and http://bugs.gentoo.org/show_bug.cgi?id=220337 ,
related Mono bugs are
https://bugzilla.novell.com/show_bug.cgi?id=471379 and
https://bugzilla.novell.com/show_bug.cgi?id=439086

Maintainer asked how can he test his patches to Mono build system. I wrote:

If you want to check whether your patch against mono doesn't break anything:
- extract a mono tarball (or do an svn checkout, but this may be broken before
you touch it ;-) )
- ./configure
- make check
If it passes, apply your changes and 'make check' again. If it doesn't pass
before your changes - apply them anyway and check if it still breaks in the
same place as before.


---------- Forwarded message ----------
From: Rafał Mużyło <galtgendo at gmail.com>
Date: Fri, May 8, 2009 at 7:35 PM
Subject: About gentoo bug 257313
To: Leszek Ciesielski <skolima at gmail.com>


I think I'll take you on that offer.
Well, that "testsuite" seems broken.
This is the line from /proc/cpuinfo:
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up
And here's what's mono doing during 'make check':
Test run:
image=/var/tmp/portage/dev-lang/mono-2.4/work/mono-2.4/mono/mini/basic-float.exe,
opts=sse2

Unhandled Exception: System.ExecutionEngineException: SIGILL
 at Tests.test_0_conv_i () [0x00000]

That's perfectly correct, this processor does not have sse2 support.
So, why does that "testsuite" tries running that test ?

BTW, I just done libgdiplus.
It does not seems to have a testsuite
and most of those test programs seem not working:
three produce no output and testgdi tries loading
non-existing files (test.png, etc.).
Just for the kicks, I've removed -fno-strict-aliasing
from that ebuild, there was only one file with a warning,
so I wrote a small patch, that seems correct to me,
that fixes that problem. Though, I'm not sure,
if it's really correct.