[EMAIL PROTECTED] wrote: >> >> So we need to know whether the right pool file is found. What is the >> output of >> >> kpsewhich --progname=pdfetex pdfetex.pool >> ls -l `kpsewhich --progname=pdfetex pdfetex.pool` >> dpkg -S `kpsewhich --progname=pdfetex pdfetex.pool` >> md5sum `kpsewhich --progname=pdfetex pdfetex.pool` >> > > aach. it's picking up /usr/local/texmf which is mounted as readonly, > over NFS. (/usr/local is mounted in this way by most of our workstations.) > So this may not be a bug after all, except that it appears to me to be a > regression w.r.t. woody.
It's not a regression, it's just that you can't do that, or rather: You can do it, but if it works then only by chance. Of course it's the purpose of /usr/local/(share/)texmf to be able to override and shadow files in /usr/share/texmf, so it is of course in order that it picks up these files. It did the same in woody, but it just happened that the woody versions and the versions in /usr/local matched, whereas the sarge versions have changed. And pool files are closely associated at compile time to the binaries; that's why they are in tetex-bin, not in tetex-base. > On our remaining woody boxes I have tetex-extra > 1.0.2+20011202 installed. Is this behaviour I am seeing the result of a > fix for wrong behaviour in that older version? No, neither a wrong behaviour nor a fix. You just have to remove all *.pool files (and probably also *.tcx) from the local tree. Regards, Frank -- Frank Küster Inst. f. Biochemie der Univ. Zürich Debian Developer
