[llvm] r309515 - Revert "[AVR] Mark a failing symbolizer test as XFAIL"

[Dylan McKay via llvm-commits]

Hey David,

> Please use SVN revision numbers rather than git hashes in commits
> There's a handy utility to help with this in llvm/utils/git-svn/git-svnrevert
- it'll take a git hash to revert, but create a local commit with the SVN
revision in the commit message instead.

That sounds really handy, thanks for that

> Also, I committed a change in r309516 that'll make this test resilient to
your change. (so adding the comment lines doesn't throw off the test's
checks for line numbers, etc) So you can probably recommit it.

Cheers!

> That said, redaing the bug, XFAILing for AVR doesn't seem like the right
answer - sounds like this would be failing on any Debian machine, AVR, X86,
etc. (so XFAIL AVR would have false positives an false negatives - there
could be AVR machines where this passes, and non-AVR machines where it
fails)

I agree that it should be fixed properly, and adding XFAIL is a dirty hack,
but I don't have the time to look into it and I'm also not familiar with
the symboliser at all.




On Mon, Jul 31, 2017 at 4:10 AM, David Blaikie <dblaikie at gmail.com> wrote:

> That said, redaing the bug, XFAILing for AVR doesn't seem like the right
> answer - sounds like this would be failing on any Debian machine, AVR, X86,
> etc. (so XFAIL AVR would have false positives an false negatives - there
> could be AVR machines where this passes, and non-AVR machines where it
> fails)
>
> Sounds like the test should be fixed.
>
> On Sun, Jul 30, 2017 at 9:08 AM David Blaikie <dblaikie at gmail.com> wrote:
>
>> On Sun, Jul 30, 2017 at 8:38 AM Dylan McKay via llvm-commits <
>> llvm-commits at lists.llvm.org> wrote:
>>
>>> Author: dylanmckay
>>> Date: Sun Jul 30 08:38:07 2017
>>> New Revision: 309515
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=309515&view=rev
>>> Log:
>>> Revert "[AVR] Mark a failing symbolizer test as XFAIL"
>>>
>>> This reverts commit 83a0e876349adb646ba858eb177b22b0b4bfc59a.
>>>
>>
>> Please use SVN revision numbers rather than git hashes in commits
>> There's a handy utility to help with this in llvm/utils/git-svn/git-svnrevert
>> - it'll take a git hash to revert, but create a local commit with the SVN
>> revision in the commit message instead.
>>
>> Also, I committed a change in r309516 that'll make this test resilient to
>> your change. (so adding the comment lines doesn't throw off the test's
>> checks for line numbers, etc) So you can probably recommit it.
>>
>> (that said, this is probably a bad test - generally good to avoid running
>> the host compiler to test functionality - most of the llvm-symbolizer tests
>> rely on checked in binaries to run over)
>>
>> - Dave
>>
>>
>>> Modified:
>>>     llvm/trunk/test/tools/llvm-symbolizer/print_context.c
>>>
>>> Modified: llvm/trunk/test/tools/llvm-symbolizer/print_context.c
>>> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/tools/
>>> llvm-symbolizer/print_context.c?rev=309515&r1=309514&r2=309515&view=diff
>>> ============================================================
>>> ==================
>>> --- llvm/trunk/test/tools/llvm-symbolizer/print_context.c (original)
>>> +++ llvm/trunk/test/tools/llvm-symbolizer/print_context.c Sun Jul 30
>>> 08:38:07 2017
>>> @@ -1,9 +1,6 @@
>>>  // REQUIRES: x86_64-linux
>>>  // RUN: %host_cc -O0 -g %s -o %t 2>&1
>>>  // RUN: %t 2>&1 | llvm-symbolizer -print-source-context-lines=5 -obj=%t
>>> | FileCheck %s
>>> -//
>>> -// See PR31870 for more details on the XFAIL
>>> -// XFAIL: avr
>>>
>>>  #include <stdio.h>
>>>
>>>
>>>
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at lists.llvm.org
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170731/1f6eb52b/attachment-0001.html>