[HarfBuzz] Thai below-base normalization

[Theppitak Karoonboonyanan]

Hi,

I'm trying to typeset Patani Malay text using Thai script as guided by
the Royal Institute, and I have some problems with Phinthu-
modified consonants with below-base vowel combined.

See the sample text captured from the book here:

http://linux.thai.net/~thep/shots/20140121-patani-sample-marked.jpg

The text is also attached as testbv.txt.

With Harfbuzz 0.9.25, I get Phinthu (U+0E3A), the consonant
modifier, below the vowel, not right below the consonant as
expected:

http://linux.thai.net/~thep/shots/20140121-patani-bv-before.png

In hb-unicode-private.hh, the combining class for U+0E3A is
modified from 103 to 3. I believe there must be some reason
behind this, but it prevents some ethnic languages from being
properly rendered. AFAIK, there are some other languages that
create new consonants this way.

I try reverting this (with the attached 01_thai-ccc.patch) and get
the proper rendering:

http://linux.thai.net/~thep/shots/20140121-patani-bv-after.png

Is the change reasonable?

In case you asked for the font I use, it's here:

http://linux.thai.net/~thep/tmp/thai-ethnic/TTFGaruda.ttf

Regards,
-Thep.
-- 
Theppitak Karoonboonyanan
http://linux.thai.net/~thep/
-------------- next part --------------
?????? ?????
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 01_thai-ccc.patch
Type: text/x-diff
Size: 608 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/harfbuzz/attachments/20140121/c0c13f2d/attachment.patch>