I've spend some time on this bug and it turned out that this is a locale
problem.
cpdftocps extracts resolution info from ppd file that belongs to the
printer. It uses sed in text extraction. In tr_TR.UTF8 locale I'm using
the extracted text is not what it's supposed to be:
utku@uther:~$ sed -nre
's/^\*DefaultResolution:\s*([0-9.]+(x[0-9.]+)?).*/resolution="${resolution:-\1}"/pi'
/etc/cups/ppd/printer.ppd
resolution="${resolution:-1200x600}"on: 1200x600dpi
utku@uther:~$ LC_ALL=en_US sed -nre
's/^\*DefaultResolution:\s*([0-9.]+(x[0-9.]+)?).*/resolution="${resolution:-\1}"/pi'
/etc/cups/ppd/printer.ppd
resolution="${resolution:-1200x600}"
Naturally this breaks the rest of the script.
There are two points here.
1. If sed's behaviour on text extraction is locale dependent, then
cpdftocps should switch to C locale (or something). Since this is a
configuration file I do not think a localized way of text extraction is
necessary.
2. I do not have any idea whether it is OK for sed to extract different
text in different locales.
Point #2 is related to sed. I'll look into it and possibly fill an
upstream bug report.
I think Point #1 is not related to upstream, since cpdftocps is not
distributed in the cups tarball, but included in distributions.
--
Utku
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
