[Piglit] [PATCH] util: migrate the SUBTEST* macros to piglit-util.sh

Martin Peres martin.peres at linux.intel.com
Thu Feb 26 07:57:27 PST 2015


On 18/02/15 19:04, Jose Fonseca wrote:
> Right. I did misunderstood how this was supposed to be used.  I saw 
> piglit_report_subtest_result callers that had many partial 
> sub-conditions.
>
> But I now see that there's the test pass/fail, and then the subtest 
> pass/fail, and each can be aggregated of multiple sub-conditions, and 
> test writers usually need to do "foo = foo && condition" at both 
> levels.  You macros target the case where each subtest is one 
> condition/error exactly.

Exactly :)

>
> So I just wondered if it wouldn't be worthwhile to devise some scheme 
> that would completely eliminate the need for test writers to keep 
> track of these aggregated pass/fail variables.

That would indeed be a good idea! Possibly have 
piglit_report_subtest_result() keep track if there have been one failure 
or not and then letting the developer query this state before returning 
the right error code.

How does this sound? I could try to propose a patch for this.



More information about the Piglit mailing list