Comment # 10 on bug 977195 from Werner Lemberg

(In reply to Petr Gajdos from comment #7)
> I insist on the statement that we do not have upstream to bother.

Unfortunately, you are right, but it might be worth to open bug reports there ���
maybe someone is listening and takes action in due course.

> Anyway, I just tried a random gujarati string and, at the first my eye
> sight, it had not expose the issue in opposite to 'Rekha-Medium' string.
> Might be that just latin is broken in the fonts?

It seems so, yes.  Using FontForge or ttx, it would be very easy to remove the
broken bytecode, or maybe to remove the non-Indic glyphs altogether.