[indic/use] Move Javanese from Indic shaper to USE
authorBehdad Esfahbod <behdad@behdad.org>
Fri, 6 May 2016 14:52:27 +0000 (15:52 +0100)
committerBehdad Esfahbod <behdad@behdad.org>
Fri, 6 May 2016 14:52:27 +0000 (15:52 +0100)
commit30e6e29f0f0f77eb65bfd1c4576b38c2da555017
tree3a6afba06f743cf58a530d38cb4ce8bca864e2d9
parentc6ee5f5f06a27a6f147f63aaf51186622821f20b
[indic/use] Move Javanese from Indic shaper to USE

Fixes https://github.com/behdad/harfbuzz/issues/243

With javatext.ttf, the reodering medial Ra gets its advance width
zero'ed in Uniscribe implementation, and the font adds the advance
back.  Our Indic shaper does not do that, but USE does.  So, route
Javanese through USE.  That's what Microsoft does anyway.  Test:

  U+A9A5,U+A9BA

This also seems to fix the following sequence, and variations thereof:

  U+A99F,U+A9C0,U+A9A2,U+A9BF
src/gen-indic-table.py
src/hb-ot-shape-complex-indic-table.cc
src/hb-ot-shape-complex-private.hh
test/shaping/texts/in-tree/shaper-indic/south-east-asian/script-javanese/misc.txt