From: Antoine Leca (Antoine10646@leca-marti.org)
Date: Thu Jun 02 2005 - 08:07:53 CDT
On Thursday, June 2nd, 2005 11:04Z Andrew West wrote:
> YNLCserif does display <0915 09BE> and other Devanagari/
> Bengali hybrids correctly (irrespective of which version
> of Uniscribe is used).
I just downloaded the version available from this page (dated september
30th, 2003), and it does not have any OpenType table for either Devanagari
or Bengali (or anything beyond Latin). This means that in effect, it does
not uses Uniscribe for rendering. What you are seeing is a simple
juxtaposition of the two glyphs, which "looks" correct in this case; Just
try <U+0915, U+09BF>, or even <U+0915, U+093F>, to see why this is fully
correct.
The problem with Uniscribe is that it does not expect the sequence <U+0915,
U+09BE>, so it considers the second to be an incomplete syllable of its own,
ill-formed so it adds a (here spurious) U+25CC glyph in between (or at the
right of uni09BF if you try it).
> However, I'm not sure that Code2000's failure to ligate 0915 and 09BE
> should be considered a problem that needs fixing.
Agreed. And similarly for Uniscribe, as I wrote.
'pôs, 'pouvez fumer ;-).
Antoine
This archive was generated by hypermail 2.1.5 : Thu Jun 02 2005 - 08:09:32 CDT