Re: [Tinycc-devel] Suggestions for test2/* tests

[Milutin Jovanović]

1) tests2/LICENSE specifies this. Did you miss this, or did I not understand your comment?

2) Good point. "make test" without failures should be a baseline for a release. However I thought I already did that (mostly). Nevertheless, I further adjusted the tests, both tests and tests2, and now "make test" passes on my osx and linux machines. I need some help with Windows tests on the other hand... (see below)

Quick status:

OSX: all tests that make/use executable files fail due to lack of Mach-O file support. 

Windows: I rarely use Windows, even less often MinGW. I had to disable many tests for it to 'pass' without failures. However this can easily be my lack of experience with the platform. I would greatly 
appreciate if someone can try other tests on Windows and report their results. (you need to enable tests in the test/Makefile and tests2/Makefile)

Linux: seems to be the most stable platform, and most tests pass. I use ubuntu, so if your results differ please give me some details. My (virtual) machines are:

Ubunty 32 bit:
$ uname -a
Linux tools-lin1 3.0.0-16-generic #29-Ubuntu SMP Tue Feb 14 12:49:42 UTC 2012 i686 i686 i386 GNU/Linux
$ gcc --version
gcc (Ubuntu/Linaro 4.6.1-9ubuntu3) 4.6.1
Copyright (C) 2011 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Ubuntu 64 bit:
$ uname -a
Linux tools-lin2 3.0.0-12-server #20-Ubuntu SMP Fri Oct 7 16:36:30 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux
$ gcc --version
gcc (Ubuntu/Linaro 4.6.1-9ubuntu3) 4.6.1
Copyright (C) 2011 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.


Miki.

On 27 June 2012 08:51, Thomas Preud'homme <address@hidden> wrote:

Hi Jovanovic,

I've just seen your commit to add more tests in the test2 directory and I have
2 suggestions:

1) it might be good to indicate the license (BSD 2-clauses if I'm right) in a
file or add the copyright notice at the top of the files.

2) as you say in the commit message, one of the purpose of tests is to avoid
regressions. Thus, it might be good for tests which fails on some architecture
to disable them on these architectures since they are not regressions and a
failure might prevent someone from commiting although he didn't break
anything. In Debian for example the package is not uploaded if the testsuite
fails.

Best regards,

Thomas Preud'homme