[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#393189: debianize-updmap not used properly



Package: texlive-base-bin
Version: 2005.dfsg.2-1
Severity: normal

Hi Norbert,

Looking at updmap from texlive-base-bin_2005.dfsg.2-1_i386.deb, I see
that debianize-updmap is sourced *before* the original definition of
syncWithTrees(). Therefore, this function won't be overloaded by the
safe version shipped in debianize-updmap.

Frank solved this in tetex-bin by moving the line that sources
debianize-updmap right after the original definition of syncWithTrees().
This is correct, therefore I didn't change that. But personally, I'd
have moved the line that sources debianize-updmap much further, that is,
after the definition of main(), i.e. after all function definitions in
updmap, so that later additions to debianize-updmap don't make it
necessary to move the sourcing place of debianize-updmap further again
(and to avoid that one forgets to move it, which would cause the new
functions not to be overloaded---silent failure, baaaad).

Also, I note that texlive-base-bin depends on texlive-common
(>= 2005.dfsg.2-1), which in turn depends on tex-common (>= 0.33).
Normally, texlive-base-bin should itself depend on tex-common (we don't
"trust" transitive dependencies), but since you control texlive-common,
I believe that's OK.

However, you should be aware that the correct behaviour of
debianize-updmap with respect to --syncwithtrees and --edit is only
present in tex-common (>= 0.34). I won't pretend, though, that this
warrants a dependency on that version, since obviously no maintainer
script should use any of these options. If a user installs
texlive-base-bin with tex-common 0.33 and hits the bug with
--syncwithtrees or --edit, then he's simply experiencing a bug that is
fixed in a later version of tex-common.

-- 
Florent



Reply to: