Re: [E-devel] [PATCH] evas BMP loader - fix for 32-bit BMP files saved by GIMP

[Carsten H. (T. R. <ra...@ra...>]

On Sun, 15 Aug 2010 11:25:51 +0200 vt...@un... said:

> Quoting "Carsten Haitzler (The Rasterman)" <ra...@ra...>:
> 
> > On Sun, 15 Aug 2010 07:16:20 +0200 vt...@un... said:
> >
> >> > it's probably a very good idea to have a little test suite for yourself
> >> > here. if it really matters to you - have a bunch of formats with a 
> >> bunch of
> >> > images that are typical of your usage (small, medium, large, simple,
> >> > complex etc. content) and then try a range of formats and options within
> >> > those formats and test time to load :)
> >>
> >> that's what the gstreamer project does. You can maybe ask them how 
> >> they do it.
> >
> > not sure there is a need to ask "how". it's simple. gather source data (the
> > images) - generate them or copy them from some set (use gimp or 
> > whatever). then
> > write a benchmark suite. none of it is the same as gtsreamer as their files,
> > code and infra are entirely different (if we wanted standardised 
> > video streams
> > to test on though it'd be relevant) :) but this is easy enough. hell it
> > could become part of expedite. :)
> >
> 
> they also have tests for image files, as they also have image loaders 
> (to create
> a video from a sequence of images). It's easy, sure. But as they 
> already did it,
> maybe they have some good advices.

i think we'll need more image files than they handle as we need to also handle
things like eet format etc. so we'll end up generating them anyway. that's
actually easier to do than even asking them. we have different rendering paths
- like we do header vs body load in 2 passes normally thus we need to be more
specific with what and how we test it (eg test a series of header loads, cached
header loads, force data loads, cached data loads, for various images from
icon-sized, to medium, to big. gst mostly needs just large images with no alpha
channels etc. - very different paths i'd say). gst has valid stuff - but it'd
not be relevant to us. well not unless its for bizarre formats where we have to
test all sorts of corner cases and WE wrote the decoder. for us thats bmp,
tga, xpm and pmaps(ppm). the rest is 3rd part or out-of-evas libs (eet for
example). we'd pass more bmp tests by far than gst right now :) so i'd rather
find more expansive test suites if anything.


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    ra...@ra...