On Mon, May 31, 2010 at 10:47 PM, <paul.sz...@sydney.edu.au> wrote: > Regardless, this bug (#584026: printconf) is hardly "done".
Yes, it is, because it doesn't apply to printconf. foomatic-filters is the only Foomatic-related package that is affected by calling gs directly, and I assume your mass-filed bug reports have hit that one too. If they haven't, you can reopen and reassign this one (or one of the umpteen others that you mass-filed without checking if gs was even called by anything in the package) to foomatic-filters. As for foomatic-filters itself: the only files specified on the command line are /dev/fd/0 and /dev/fd/3, and gs is called with -DPARANOIDSAFER (which appears to be equivalent to -DSAFER nowadays). That would seem to narrow the vulnerability window, assuming only files in /dev/fd could be accessed, although I suppose (although I have no idea how) there might be a way to get gs to generate something on the error or other streams that could be injected back in (e.g. one of the numbered fds). I would suggest also that this is not a Debian-specific issue in Foomatic and probably should be reported to the foomatic-devel list; upstream may prefer a different approach to fixing the issue than the one I would take. Chris -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
