[archeology] Re: [llvm] r218742 - test: XFAIL the non-darwin gmlt test on darwin

[Justin Bogner via llvm-commits]

Adrian Prantl via llvm-commits <llvm-commits at lists.llvm.org> writes:
> Hello [random folks who touched this file in the past couple of years]!
>
> Does anyone understand how this test is supposed to work?
>
> The test "test/DebugInfo/Generic/gmlt.test” is building and
> FileCheck’ing "test/DebugInfo/Generic/Inputs/gmlt.ll” but is XFAILing
> it under Darwin.
> This is odd, but maybe ok.
>
> But then "test/DebugInfo/Generic/Inputs/gmlt.ll” also has its own RUN line:
>
>   ; RUN: %llc_dwarf -O0 -filetype=obj < %s | llvm-dwarfdump - | FileCheck %s

This file is in Inputs, so lit ignores it. The run line here doesn't do
anything.

> and a couple of darwin-specific checks:
>
>   ; DARWIN:  DW_TAG_subprogram
>   ; DARWIN-NOT: DW_TAG
>   ...
>
> Now, as far as I can tell, there is no FileCheck invocation that uses
> a --check-prefix=DARWIN to be found. And changes to the DARWIN check
> lines seem not to disturb a "check-llvm”.

Looks like this file is duplicated into both DebugInfo/Inputs/gmlt.ll
and DebugInfo/Generic/Inputs/gmlt.ll, and the test under X86 invokes the
DARWIN flavoured test on DebugInfo/Inputs/gmlt.ll. The whole structure
is quite confusing.

We should either (a) take these both out of Inputs and replace the X86/
and Generic/ tests with complete .ll files instead of strange includes,
or (b) make both tests refer to the same Inputs/*.ll file.

> - Do we want this test to be executed twice?
> - Should we re-enable it on Darwin?
>
> -- adrian
>
>
>> On Sep 30, 2014, at 10:45 PM, Justin Bogner <mail at justinbogner.com> wrote:
>> 
>> Author: bogner
>> Date: Wed Oct  1 00:45:45 2014
>> New Revision: 218742
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=218742&view=rev
>> Log:
>> test: XFAIL the non-darwin gmlt test on darwin
>> 
>> r218702 disabled a -gmlt optimization for darwin, but this means the
>> non-darwin test isn't working there anymore.
>> 
>> Modified:
>>    llvm/trunk/test/DebugInfo/gmlt.test
>> 
>> Modified: llvm/trunk/test/DebugInfo/gmlt.test
>> URL:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/test/DebugInfo/gmlt.test?rev=218742&r1=218741&r2=218742&view=diff
>> ==============================================================================
>> --- llvm/trunk/test/DebugInfo/gmlt.test (original)
>> +++ llvm/trunk/test/DebugInfo/gmlt.test Wed Oct  1 00:45:45 2014
>> @@ -1,2 +1,5 @@
>> ; REQUIRES: object-emission
>> ; RUN: %llc_dwarf -O0 -filetype=obj < %S/Inputs/gmlt.ll |
>> llvm-dwarfdump - | FileCheck %S/Inputs/gmlt.ll
>> +
>> +; There's a darwin specific test in X86/gmlt, so it's okay to XFAIL
>> this here.
>> +; XFAIL: darwin
>> 
>> 
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits