[tex-live] Bug#425085: texlive-doc-en: catalogue entries with dangling links

Robin Fairbairns Robin.Fairbairns at cl.cam.ac.uk
Thu Jun 21 01:00:52 CEST 2007

Karl Berry <karl at freefriends.org> wrote:

>     Doesn't look straightforward to fix - any ideas?
> We will probably reinstate the Catalogue in TL, and when we do, I guess
> we should not just copy the files, but set up some kind of "conversion"
> process.  Ugh.

the catalogue as you see it on ctan or sarovar is the result of
processing the xml files that make up the repository.

since those xml files contain the names of tpm files (for those for
which i know the correspondence).  for example:

<entry datestamp='$Date: 2007-06-20 23:40:03 +0100 (Wed, 20 Jun 2007) $' modifier='$Author: robin $' id='footmisc'>
  <texlive location='footmisc'/>

[the id is the name of the xml file (and hence the name of the converted
html file); the texlive location is the name of the tpm file.]

given that info, one could in principle match things up, presumably?

but before we even start, we need to know what we want the links to
point to.  suppose ctan has pdf docs and texlive has .dvi ... or, as in
the case of footmisc, ctan has pdf and texlive has no docs at all?

> Does MiKTeX contain the Catalogue?  If so, are the links properly
> transformed?  If so, maybe we could just work from that.

it doesn't, that i can see.  (there's a file called files.csv.bz2 in the
miktex package repository, that tells you every file in miktex and the
miktex package it belongs to.  i've no script that looks in all the
different formats and extracts the data.)

> If someone wants to take this up, great.

as it stands, i don't think it's do-able.  we need a definition of
what's really required.  (this comes as a surprise to me ... until this
issue arose, i had assumed it was all trivial.)

would a non-ctan html version, as you find on sarovar, help?


More information about the tex-live mailing list