[Pixman] [PATCH 1/1] Disable implementations mentioned in the PIXMAN_DISABLE environment variable.
sandmann at cs.au.dk
Sun Feb 26 14:01:15 PST 2012
Andrea Canciani <ranma42 at gmail.com> writes:
> I think that strstr might cause substrings to also disable unintended backends.
> This would already happen for "mmx" vs "arm-iwmmxt".
> Is this intentional?
> Although in this specific case, it might be ok, I believe that it
> might lead to unexpected behavior as other backends are added.
>> Indeed, this is very useful.
>> I often seen if (!disabled(envvar) && supported()) where the disabling
>> check is first in the if-statement. That's the only suggestion I have.
>> Otherwise, this looks like a good change.
> I agree that if the backend is disabled, we should avoid checking the support.
> This way, we can use disable to ensure that the backend cannot break
> the application status in any way (example: modify simd registers
> without restoring them appropriately).
I agree with both comments. New patch to follow.
More information about the Pixman