Hi Sergei,

thanks for the report.

Sergei Ivanov <[EMAIL PROTECTED]> wrote:

> Package: tetex-bin
> Version: 3.0-29
> Severity: normal
>
> On my system, the command
>   mktexpk --mfmode ljfour --bdpi 600 --mag 1+120/600 --dpi 720 lhr10
> places the resulting file lhr10.720pk in the current working directory.
> (The font lhr10 is a cyrillic version of cmr10, I think it belongs
> to tetex-extra.)
[...]
> The difference between lhr10 and cmr10 is that I have lhr10.tfm
> and some other lh* files under /var/lib/texmf (they were probably 
> created by fmtutil since I use custom formats with these fonts).

I don't think fonts should ever be created there.

> "rm -rf /var/lib/texmf/fonts" fixes the problem.

Hm, I tried to reproduce this, but I did it as root and the resulting
file was created in /var/lib/texmf/fonts were I put the source files,
too.  From this alone I would conclude that the scripts try to put the
files in the same tree as the sources, and use the current directory as
fallback if the normal target directory is not writable.  I mention this
since it means there might be a reason for this behavior...

> I tried to debug it and found that one of the scripts (maketexnam)
> finds lhr10.tfm under /var/lib/texmf but fails to recognize that
> /var/lib/texmf is a "system" tree.

That's correct.  mktexnam asks kpsewhich for SYSTEXMF, which is defined
in /etc/texmf/texmf.d/05TeXMF.cnf:

% The system trees.  These are the trees that are shared by all the users.
SYSTEXMF = $TEXMFLOCAL;$TEXMFMAIN;$TEXMFDIST

According to this description, this should include TEXMFSYSVAR and
TEXMFSYSCONFIG.  However, it's not as simple as just adding these trees
to SYSTEXMF.  In mktexnam, MT_VARTEXFONTS is explicitly added to the
trees in SYSTEXMF.  This variable used to point to /var/cache/fonts, now
it points to TEXMFVAR/fonts, with /tmp/IforgotWhat as fallback.
/var/lib/texmf/fonts is and never was supposed to contain pk fonts, if I
remember correctly.


So maybe we should rather find out how these fonts got there.  If it is
something that might have happened to more users, we may need to cater
for it, if it's just a special case on your system, maybe not.

Regards, Frank
-- 
Dr. Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)

Reply via email to