lohit-devanagari is almost ready for beta release after resolving number of issues reported at git [4]
I am happy while testing lohit2 devanagari we are not only fixing issues from fonts side but we found interesting bugs from application side as well.
Earlier it was from Harfbuzz NG side. [1] This bug sometime very difficult to identify since even when expected ligature does not come it looks like there is ligature. During thorough testing with hb-shape and by looking at glyph names we identified yes this is problem. Then we tested with other fonts around and no one was giving expected results.
Thanks to Behdad for quickly fixing this it made life easy for doing Beta release.
Now few days back we identified one more bug from Libreoffice [2] This is also tricky one. Surprisingly this was working well with NotoSans Devanagari so one might think it is font issue as it work well with other fonts.
But again with thorough analysis we found that changing glyph class of first-i-matra to base-class doesn't solve the issue but create other problem for forming reph ligatures. [5]
This time survivor is Khaled, good to see he committed the fix today itself. Though my testing is remainign but he resolved the root cause i think.
From Fedora side, as Fedora is the always leader for quick fixes from upstream. Caolan again proved that by providing build [3] for fix today itself. Thats simply great !!!
I am sure over the time we will done with Lohit2. We will definitely identified major issues in rendering stack of opensource and will have the fixes also ready with excellent people around. I am reoommending community member to participate in lohit2 testing since this will help for long time to have perfect fonts and Open type layout shapers around.
Though this making bit delay in lohit2 releases but i think perfection comes first :)
1. https://bugs.freedesktop.org/show_bug.cgi?id=69266
2. https://bugs.freedesktop.org/show_bug.cgi?id=71080
3. https://bugzilla.redhat.com/show_bug.cgi?id=1025201
4. https://github.com/pravins/lohit/issues
5. https://bugs.freedesktop.org/show_bug.cgi?id=70968#c9
I am happy while testing lohit2 devanagari we are not only fixing issues from fonts side but we found interesting bugs from application side as well.
Earlier it was from Harfbuzz NG side. [1] This bug sometime very difficult to identify since even when expected ligature does not come it looks like there is ligature. During thorough testing with hb-shape and by looking at glyph names we identified yes this is problem. Then we tested with other fonts around and no one was giving expected results.
Thanks to Behdad for quickly fixing this it made life easy for doing Beta release.
Now few days back we identified one more bug from Libreoffice [2] This is also tricky one. Surprisingly this was working well with NotoSans Devanagari so one might think it is font issue as it work well with other fonts.
But again with thorough analysis we found that changing glyph class of first-i-matra to base-class doesn't solve the issue but create other problem for forming reph ligatures. [5]
This time survivor is Khaled, good to see he committed the fix today itself. Though my testing is remainign but he resolved the root cause i think.
From Fedora side, as Fedora is the always leader for quick fixes from upstream. Caolan again proved that by providing build [3] for fix today itself. Thats simply great !!!
I am sure over the time we will done with Lohit2. We will definitely identified major issues in rendering stack of opensource and will have the fixes also ready with excellent people around. I am reoommending community member to participate in lohit2 testing since this will help for long time to have perfect fonts and Open type layout shapers around.
Though this making bit delay in lohit2 releases but i think perfection comes first :)
1. https://bugs.freedesktop.org/show_bug.cgi?id=69266
2. https://bugs.freedesktop.org/show_bug.cgi?id=71080
3. https://bugzilla.redhat.com/show_bug.cgi?id=1025201
4. https://github.com/pravins/lohit/issues
5. https://bugs.freedesktop.org/show_bug.cgi?id=70968#c9