Thomas Ruedas ruedas@geophysik.uni-frankfurt.de
Mon, 13 Mar 2000 17:26:28 +0100

>Value of <HANDLE> construct can be "0"; test with defined() at ./texdoctk line >65535.
>Value of <HANDLE> construct can be "0"; test with defined() at ./texdoctk line >65535.
>Value of <HANDLE> construct can be "0"; test with defined() at ./texdoctk line >65535.
Which Perl version do you have? I got this crap as well on my Linux box
at home (Perl v.5.004, IIRC), but didn't notice any particular problems;
note that I have the -w option in the first script line. I can't test
now if it would disappear if you take it away, because with the slightly
newer Perl here (AIX, Perl v.5.005_03) it doesn't show up.
>Read on closed filehandle <UPDATES> at ./texdoctk line 56.
>Use of uninitialized value at ./texdoctk line 57.
Lines 54-56 are
    my $updates=join('/',($texmfmain,"updates.dat"));
so updates.dat is expected to be in $TEXMFMAIN. Could it really be in
another place or be absent at all? I don't know if you tried as root or
as a normal user; the file must be world-readable, I guess.
Please tell me about that, so that I can make a workaround.
>But I cannot see a trivial way to keep it up to date or extend the database
I can't either, but I am prepared to maintain databases for the original
releases of new teTeX texmf versions as well, of course. Maintaining the
database for local specials naturally is the duty of the local sysadmin,
but just adding a line if you install a *new* package should not be such
a tremendous effort.
I don't think there can be a trivial way, because the frontend needs a
kind of slogan for the user to know what a package is about, and also
some keywords maybe, for the search.
>would it be better NOT to give the path of files, but just let
>kpsewhich find them?
Running kpsewhich when a file is selected, you mean? However, I am not
sure if kpsewhich would find the correct result in each case (although
in most cases, it would). But there are some files or at least
possibilities of file names such as index.html or README which are or in
future versions might be not unique.

Thomas Ruedas
Institute of Meteorology and Geophysics, J.W.Goethe University Frankfurt
e-mail: ruedas@geophysik.uni-frankfurt.de