Please do not reply directly to this email. All additional comments should be made in the comments box of this bug report.
Summary: Review Request: dvipost - latex post filter command
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=190071
------- Additional Comments From mpeters@mac.com 2006-04-27 11:06 EST ------- (In reply to comment #11)
Something like:
%{!?_texmf: %define _texmf %(eval "echo `kpsewhich -expand-var '$TEXMFMAIN'`")}
Well, except that defining _texmf would in this case cause an error as well unless configure was patched to take it as an arguement (which I don't think is necessary) - I guess in this case expecting to support building against modified tetex environments might be a bit much because of the upstream configure script which looks for a specific file and doesn't take a texmf as a switch.
That makes sense but then it would imply to Require: tetex-doc. That would mean that a 40 KB package could potencially require an 100 MB package. I don't think this is worth it. :-)
/usr/bin/texdoc is owned by tetex. The potential problem is who owns the directories within the tex documentation tree if tetex-doc isn't installed - but other packages just own it themselves.
Since it is just the man page, and available as a man page, it isn't that big of a deal.
Actually I think that dvipost requires a tex installation, there is nothing exclusive from tetex. That was the reason why I have proposed dvipost and not tetex-dvipost.
If you feel strongly about this I will rename it.
On fedora - tetex is what provides tex. There are other examples of this (in core)
[mpeters@atlantis Desktop]$ rpm -qf /usr/bin/dvips tetex-dvips-3.0-17 [mpeters@atlantis Desktop]$ rpm -qf /usr/bin/xdvi tetex-xdvi-3.0-17
It also makes it a little easier to find when browsing repoview for tetex boltons.