References: <18385.42387.268037.81145@morse.mittelbach-online.de>
            <85wsoeto1k.fsf@lola.goethe.zz>
            <18385.42387.268037.81145@morse.mittelbach-online.de>
            <5.1.0.14.0.20080310115827.01e05a70@pop3.web.de>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.0.60 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Message-ID:  <868x0qk89o.fsf@lola.quinscape.zz>
Date:         Mon, 10 Mar 2008 14:39:47 +0100
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From: David Kastrup <dak@GNU.ORG>
Subject: Re: A really, really bad bug.
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
In-Reply-To:  <5.1.0.14.0.20080310115827.01e05a70@pop3.web.de> ("Uwe
              =?iso-8859-1?Q?L=FCck=22's?= message of "Mon, 10 Mar 2008
              12:19:12 +0100")
Precedence: list
Status: R

Uwe Lück <uwe.lueck@WEB.DE> writes:

> At 21:44 07.03.08, you wrote:
>>Frank Mittelbach <frank.mittelbach@LATEX-PROJECT.ORG> writes:
>>
>> > well, you seem to be the first person getting into trouble with this,
>> > but that doesn't mean one shouldn't fix it.
>>
>>5000 labels is easy to reach if you are using the label mechanism
>>extensively.
>
> This is real life with critical editions in the "classical" style
> where footnotes don't refer by footnote marks but by line numbers.
> ednotes.sty originally used three labels for one footnote
> and broke with, say, 400 pages. Well, the problem was the number
> of strings already, but you see that so many labels are what
> some people really need. I changed the label-testing to use only
> one label per note, but it is not very stable. Think of a volume of
> 500 pages (I am working at something like that) and that ten notes
> per page are not unusual ...

I have added a patch to my local version of perpage.sty.  I think I
should be able to upload it to the CVS today (and send a copy to this
list) once I have conditionalized the patch in order to make sure that
it is only applied when \@testdef leaves a permanent \relax record.

This conditionalization means that it would be a good idea to solve this
bug in LaTeX via the additional grouping that is already used when
creating the label (so that the patch will then get deactivated).  Since
this is already done in the \@newl@bel (sp?) code, too, I think this is
the most consistent way of addressing the problem.

I'd fix this in the LaTeX source myself too, but as I said: it will
break Babel (which explicitly _checks_ for the old definition), and I
have not heard any proposal how we can propagate the fix there.

And I would really want to avoid using \AtBeginDocument in the kernel
for delaying the change of definition until after Babel is loaded.

-- 
David Kastrup