[XeTeX] xdvipdfmx patch

Alexey Kryukov anagnost at yandex.ru
Sun Dec 15 22:34:52 CET 2013

Hi all,

I have implemented a patch intended to improve searching and copying
text from pdf files produced by xdvipdfmx.

With this patch xdvipdfmx will attempt to extract Unicode values from
glyph names for PUA glyphs and alphabetic presentation forms just like
it currently does for unencoded glyphs (the PUA Unicode value is still
used as a fallback choice). For those TTF fonts which have no glyph
names in their 'post' table there is an attempt to recover the source
Unicode by reverting the GSUB rules. Thus is is now possible to enable
typographic features in fonts which ship with MS Windows (e. g. Times
New Roman or Palatino Linotype) without breaking the document

I see there is no bug tracker for texlive, so I am sending this patch
here hoping that somebody can test and apply it.

BTW I have noticed the following bug in the current texlive source
tree: xdvipdfmx is linked with cidtype0.o for dvipdfmx, which causes it
to fail on any attempt to load an OT-CFF font. This can be fixed by
adding cidtype0.[ch] to dvipdfmx_SOURCES in Makefile.am (and rerunning

Alexey Kryukov <anagnost at yandex dot ru>

Moscow State University
Faculty of History
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xdvipdfmx-pdfcmap.patch
Type: text/x-patch
Size: 16626 bytes
Desc: not available
URL: <http://tug.org/pipermail/xetex/attachments/20131216/9ba0b132/attachment.bin>

More information about the XeTeX mailing list