Frank Küster <[EMAIL PROTECTED]> wrote: > On the other hand, I'm not sure how to fix this. There's a reason why > the update-* scripts fail (with an understandable error message) when > the basic file is missing - so this shouldn't be changed. Should we > really ignore all errors of update-* in "postrm remove"?
Tough question. It would certainly "fix" the error... apparently, but such a fix may lead to later breakages, more difficult to analyze. If you choose to ignore the errors, I'd suggest not to ignore them altogether, but actually perform the update-* and updmap-sys/fmtutil-sys calls in postrm scripts only when tex-common is configured (by checking for the existence of 00tex.cnf, or by using dpkg-query?...). If tex-common is not configured, the TeX system is in such a bad state that we cannot do much to help the packages getting removed properly. Granted, supposing the above is implemented: if the package to remove was texlive-lang-french, it could be removed from dpkg's POV while my formats would still have french hyphenation patterns. But I don't think it's that bad, because the TeX system *is* broken at that time. Only when tex-common is configured, may the TeX system be in a sane state. Uuups. I wrote that under the assumption that tex-common tries to run updmap-sys and fmtutil-sys after the update-* commands, but it doesn't. Therefore, in my example, my formats would still have french hyphenation patterns installed even after tex-commmon is configured again, with texlive-lang-french removed. Baaad. My conclusion: if these errors are avoided as described above, then tex-common should also try to call create_tetex_formats() and create_fontmaps() when configured (which, in case of errors, would probably spit out more confusing messages...). -- Florent
