[XeTeX] XeLaTeX PDF Glitch with Serbian Glyphs

Georg Duffner g.duffner at gmail.com
Wed Sep 18 11:19:48 CEST 2013


Am 17.09.2013 16:55, schrieb Alessandro Ceschini:
> Hello Ross,
>
> I attached the whole output, not just the pdf, in the even that you wish
> to check the .tex file as well. I've also attached the output, that is
> the pasted text from the .pdf as it appears on LO. The font is always
> the same, FreeSerif, bundled with Ubuntu.

The locl feature of FreeSerif is a bit problematic (but I’m not sure if 
it’s the only issue here, not having enough time to test thoroughly).

FreeSerif (I tested with the version on CTAN) is a bit inconsistent in 
its glyph naming. In the attached screenshot you see the incriminated 
locl feature of FreeSerif Italic which defines

becyrillic  → serb.be
tecyrillic  → serb.alt_te
decyrillic  → cyrillic.alt_de
gjecyrillic → cyrillic.macedon_gje
pecyrillic  → uni04E3
gecyrillic  → imacron

and in salt (not in the screenshot), there is

shacyrillic → serb.alt_sha

The convention on glyphnaming is, that the base name is before the dot 
and the suffix specifies the variant 
(http://sourceforge.net/adobe/aglfn/wiki/AGL%20Specification/). Thus 
it’s expected that the glyphs be named becyrillic.serb, decyrillic.alt etc.

The replacements of pecyrillic and gecyrillic however are buggy. 
pecyrillic is replaced by uni04E3 which is the cyrillic i with macron 
and gecyrillic is replaced with latin i with macron. When copy/pasting 
from a pdf they may get pasted exactly as these i’s with macra (which in 
my superficial tests with XeLaTeX actually happens). This is the same 
issue as with LinLibertine which I described in my last mail.

Best regards,
Georg
-------------- next part --------------
A non-text attachment was scrubbed...
Name: FreeSerifItalic_locl.png
Type: image/png
Size: 36858 bytes
Desc: not available
URL: <http://tug.org/pipermail/xetex/attachments/20130918/267195c0/attachment-0001.png>


More information about the XeTeX mailing list