Steve Langasek wrote: > On Tue, Aug 23, 2005 at 07:24:51PM +0200, Matthijs Mohlmann wrote: > >>tag 318384 + patch >>thanks > > >>First of all: > > >>texhash: /usr/local/share/texmf/ls-R: no write permission. Skipping... >>texhash: /usr/local/lib/texmf/ls-R: no write permission. Skipping... >>texhash: Updating /var/lib/texmf/ls-R... >>texhash: Updating /var/lib/texmf/ls-R-TEXMFMAIN... >>texhash: Updating /var/cache/fonts/ls-R... >>texhash: Done. > > >>The make install target is doing running texhash by it's own while this >>should be done into the postinst / postrm of a debian package. > > >>Secondly it tries to update the font map files in the install target >>(especially the install-data-hook target): > > >>using config file /var/lib/texmf/web2c/updmap.cfg >>updmap: The options --enable and --disable work differently on Debian >>systems!. >>updmap: . >>updmap: You cannot use them to permanently enable or disable Map files. >>Instead,. >>updmap: you have to use the mechanism described in. >>updmap: /usr/share/doc/tetex-bin/README.Debian.gz. >>updmap: . >>updmap: If you don't know which lines to add to the files in >>/etc/texmf/updmap.d/,. >>updmap: press ENTER now and take the resulting line from. >>updmap: /var/lib/texmf/web2c/updmap.cfg. >>updmap: . >>updmap: Press ENTER to continue anyway, or Ctrl-c to quit.. > > >>Attached is a patch that removes the install-data-hook target from the >>install target. > > > This really seems like the wrong fix, IMHO; I can understand that updmap > in Debian may not have the same behavior as upstream's, but I don't > understand why it should be gratuitously different and break when > invoked non-interactively, as from a build script. >
Then there is a bug in the debian updmap script / binary. But still the updmap and texhash should not be called by the make install. This can break a debian system IMO. (Maybe not a buildd, but a person who wants to customize the source) Look also at the files it is trying to affect, texhash is trying to do something in /var/lib/... and the updmap is trying to do something in /etc. Also the postinst and postrm have a call to updmap-update and texhash which actually doing the same when installing from source. So there is IMO nothing wrong with this fix, just some redundancy.
signature.asc
Description: OpenPGP digital signature
