[Bf-committers] using checksum instead of bundling images for regression tests?

Daniel Fairhead bf-committers@blender.org
Sat, 22 May 2004 00:10:58 +0300

> > Is the image output for renders consistent enough that we could checksum the 
> > images to see if they rendered properly?
> No.
> And strange suggestion anyway... do you think testers are not
> smart enough to decide if the render is correct or not by just
> comparing it? ;)

No offence intended, but many people aren't _aware_ _enough_
to notice rendering errors. For instance, certain types of jaggies,
incorrect jitter, overly bright reflections, missing textures (if they 
aren't expecting one there), etc.

If the images were consistant enough, also, the rendering could be
scriptified, for simpler testing, calling the "blender -B" from a bash-script,
and the checksumming the results, which would be easier (allthough
less fun?)