Bug#227398: tetex-extra: Also destroys listoffigures and listoffigures
James McCaw <jamesm@physics.unimelb.edu.au> schrieb:
> My only comment is that a default broken link is clearly identifiable
> with hyperref, while a default overfull hbox could come from any
> number of packages or user errors.
Hm. I would argue just the other way round: A user unfamiliar to
hyperref will have a hard time finding out why the advertised behavior,
the creation of hyperlinks, does not work in the toc/lof/lot. Especially
since the hyperref documentation is not one of the best.
On the other hand, if adding hyperref completely destroys the word wrap,
it is quite clear that he must have done something wrong in the handling
of this package, and can start googling or looking into the FAQ.
> The thing that took me a long time
> to debug was establishing whether it was my fault (a coding error) or
> a package issue. Once I narrowed the behaviour down a package, I then
> had to delete packages until I found it was hyperref.
If an error occurs at a specific place in the typed text, it makes sense
to first narrow the document step by step until only the problematic
part is left, and to start removing packages afterwards. But if the
problem is in the toc/lot/lof, I do not understand why you don't start
by trying out a sample document with one figure with a copy of a real
caption. Then you could have come to narrowing it down to a specific
package quite soon.
> Apart from my first suggestion, which seems to not be a good idea, I
> really can't think of a clean/sensible/noticeable way to warn a user
> of this without the user already knowing that the package hyperref is
> the issue. I'll give it some thought...
I fear the ultimate solution would be to write a new documentation for
hyperref. Properly sectioned (user, expert, developer,...), not
dispersed over a couple of documents, and all that. But this won't
happen until somebody starts to work together with Heiko Oberdiek,
hyperref's maintainer, and does most of the writing.
Regards, Frank
--
Frank Küster
Inst. f. Biochemie der Univ. Zürich
Debian Developer
Reply to: