On Tue, 27 Nov 2012 13:44:14 -0500, Alexei Podtelezhnikov said: >Glyph rendering quality is in the eye of the beholder. What somebody >says is blurry is smooth in the eyes of the other. How do you suggest >that we automate that?
Yes, that's tricky. But that could be delayed until later. Since there is *no* testing currently, it would be an improvement to at least have tests that generate *some* output. Fuzzy comparison can come later. You could at least catch regressions that cause tests to crash. The VTK project, which uses freetype, has tests that generate images and then compare images. If less than x pixels are different than one of several 'correct' images, the test fails. Ex: <http://open.cdash.org/testDetails.php?test=167853581&build=2684371> <http://open.cdash.org/testDetails.php?test=167853492&build=2684371> On Tue, 27 Nov 2012 10:53:18 -0800, Brian Stell said: >So, crowdsourcing the test systems / build-bots? And the aggregating the >results? Exactly. Sean _______________________________________________ Freetype-devel mailing list Freetype-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/freetype-devel
