The cmap code in graphite (bug 631479) doesn't understand format-14 cmaps and so doesn't handle IVS sequences correctly.
Used the existing IVS reftest and added a dummy Silf table to force the font through the graphite shaper (assuming it's enabled).
There are no known fonts that require both format14 and graphite. I'm happy to work on this, but I do not see this as blocking bug #631479. Please could you remove the block. It'll get sorted out and appear in the code soon with no API change. We can sort this out in trunk. Please could you include a binary patch of your sample font for testing, I can't build a file from a git patch against a git repo I don't have.
Sorry. Misread the bug. This is a follow on not a blocker. I'll add it as something to do sometime. But a proof of need font would be useful.
Martin, the reftest includes a font with the graphite/format14 combination.
yes but that particular file consists of a git diff against a repo I don't have which involves a binary git diff format that I don't seem able to unpack into a file. Could you just put up the single file from that reftest as a binary file. TIA
This is just our usual patch format which you can apply against a firefox source tree. After you've set up mercurial and pulled sources, you can import that patch above: hg qimport -P graphite-ivs-reftest.patch The reftest will be in layout/reftests/font-face and the font in layout/reftests/fonts. If you're still having trouble with this I can send you the files directly.
I think Martin's concern is that it's rather burdensome to have to pull the entire mozilla tree and apply the patch just to get hold of one (tiny) test font. So I'm uploading the .ttf file as a separate attachment to make it more easily accessible.
You need to log in before you can comment on or make changes to this bug.