[compiler-rt] r238874 - Tests: disable test of /proc filesystem on Darwin.

[Jonathan Roelofs]

On 6/5/15 12:32 PM, Evgeniy Stepanov wrote:
> On Fri, Jun 5, 2015 at 11:23 AM, Duncan P. N. Exon Smith
> <dexonsmith at apple.com> wrote:
>>
>>> On 2015-Jun-05, at 11:11, Evgeniy Stepanov <eugeni.stepanov at gmail.com> wrote:
>>>
>>> It looks like FreeBSD may have /proc, so moving the test to Linux
>>> would be bad.
>>
>> Although it might be disappearing there, based on this ancient
>> reference Wikipedia has on its procfs page:
>>
>> http://lists.freebsd.org/pipermail/freebsd-fs/2011-February/010760.html
>>
>>> You could add XFAIL: darwin, or, in this particular
>>> case, simply exit(0) if the file in /proc could not be opened.
>>
>> Anyone know lit well enough to add a `procfs` feature?  Seems like
>>
>>      REQUIRES: procfs
>>
>> would be the clearest answer.
>>
>> I don't like `XFAIL`, since this isn't really a test failure that
>> we're going to try to fix eventually -- it's a feature that Darwin
>> doesn't have and isn't planning to grow.
>
> Yes, unfortunately lit does not support excluding a single test based
> on certain conditions (smth like REQUIRES: NOT darwin, or UNSUPPORTED:

`UNSUPPORTED: darwin` is the spelling that libcxx and libcxxabi use for 
this sort of thing.  Not sure if compiler-rt's lit.cfg supports that.


Jon

> darwin would be useful, maybe even allowing complex boolean
> expressions). Also, in this case "procfs" feature would not be enough
> because the condition should look like procfs && !android. The test is
> broken on Android for other reasons.
>
>>
>>>
>>> On Fri, Jun 5, 2015 at 11:03 AM, Tim Northover <tnorthover at apple.com> wrote:
>>>>> IMO this test should be moved to Linux/ (and duplicated to an Android/
>>>>> directory?),
>>>>
>>>> I thought that originally, but Wiki says /proc filesystems are on at least some BSDs too (though I'm not sure if they support the particular access being done here).
>>>>
>>>>> or moved to a LinuxLike/ directory, or the individual test
>>>>> file disabled on Darwin.
>>>>
>>>> I didn't know you could disable individual files, that sounds like a reasonable change if it is possible though.
>>>>
>>>> Tim.
>>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>

-- 
Jon Roelofs
jonathan at codesourcery.com
CodeSourcery / Mentor Embedded