Rumor has it that a patch exists which will fix the way that pretty much every Android reftest run fails one test or another, with one chunk of text differing in the antialiasing, or the join between two colored borders, or around the edge of a rounded corner. I want to see that patch!
probably a misunderstanding -- i do not have such a patch. However, I understand some of the failures you are seeing. On Android, we: disable color correction do not use pango (font shaping) use 16bpp not 24/32 Any of these may account for some of the issues that you are reporting.
Not sure what, but something's changed in the last two months - at the time I filed it, six or eight times a day I was opening the reftest analyzer and saying "yup, antialiasing around text" or "yup, antialiasing of rounded borders," and now I don't really remember the last time I looked at one.
Status: NEW → RESOLVED
Last Resolved: 7 years ago
Resolution: --- → WORKSFORME
Something since the last Aurora branch point - just looked at some failed runs there, and they are just like I remember things when I filed this.
I expect this was fixed by bug 637852. I know I saw some antialiasing-related reftest failures within the last week or two, though not in the past few days.
You need to log in before you can comment on or make changes to this bug.