[Piglit] [PATCH 0/7] Reducing ortho use in shader_runner

Dylan Baker baker.dylan.c at gmail.com
Tue Nov 5 10:30:11 PST 2013


sanitizing the filenames would be relatively easy. Python has a number of such 
functions builtin, it shouldn't be to hard to find one. All you'd need to do 
is hook it into Summary.generate_html(), there are a bunch of with ``open(...) 
as ...'' statements that you'd sanatize.

On Tuesday, November 05, 2013 12:16:47 PM Ken Phillis Jr wrote:
> I believe the shell scripts are pretty much in the seven patches discussed
> on this main patch details.
> 
> 
> Also, I was wondering Would it be considered brainless to simply implement
> a filename sanitizer for the HTML generator? The glean tests for example
> cause all sorts of problems on windows due to invalid characters. The
> basics of what this entails is on Wikipedia (
> http://en.wikipedia.org/wiki/Filename ) I figure that it would be
> beneficial to make sure that generated files don't try to place invalid
> characters that lead to programs failing.
> 
> 
> As a tip, Just run the glean tests on windows and then try to make a html
> output for it... It'll fail every time due to invalid characters... I have
> tested this on Windows 7 and it should also apply to any version of windows.
> On Tue, Nov 5, 2013 at 12:09 PM, Dylan Baker <baker.dylan.c at gmail.com>wrote:
> > On Tuesday, November 05, 2013 09:38:16 AM Ian Romanick wrote:
> > > On 11/04/2013 10:18 PM, Matt Turner wrote:
> > > > We'd like to shrink the window sizes of shader runner when running in
> > > > simulation, and ortho tests make it difficult to reason about reducing
> > > > the window size.
> > > > 
> > > > The following patches fix some tests and test generators to use
> > > > non-ortho projections instead. This should allow us to shrink these
> > > > tests' window sizes and execute them faster. They reduce the number of
> > > > occurrences of 'ortho' in shader_test files in piglit from 920 to 153.
> > > > 
> > > > I have some additional patches that delete these generated tests from
> > > > git and move their generators to generated_tests/, but since the
> > > > generators are written in bash (I assume?) they're really slow. I'm
> > > > not sure if it's worth it, but we would remove 60k lines of generated
> > > > code from piglit.
> > > 
> > > The bigger problem is that piglit builds on platforms that may not have
> > > bash.  Right?  We'd have to convert them to Python... and that feels
> > > like too much work.  Too bad we don't still have a bunch of interns
> > > around. :)
> > > 
> > > > (I'm not sending out patches containing the regenerated tests, for
> > > > obvious reasons)
> > > > 
> > > > Thanks,
> > > > Matt
> > > > _______________________________________________
> > > > Piglit mailing list
> > > > Piglit at lists.freedesktop.org
> > > > http://lists.freedesktop.org/mailman/listinfo/piglit
> > > 
> > > _______________________________________________
> > > Piglit mailing list
> > > Piglit at lists.freedesktop.org
> > > http://lists.freedesktop.org/mailman/listinfo/piglit
> > 
> > I'm kicking for something brainless to do, are these generators laying
> > around
> > somewhere? I'll have a look at how hard it would be to convert them to
> > python.
> > _______________________________________________
> > Piglit mailing list
> > Piglit at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/piglit
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.freedesktop.org/archives/piglit/attachments/20131105/49bc6073/attachment-0001.pgp>


More information about the Piglit mailing list