[tex-live] bibtex's max_print_line

Oren Patashnik op at CS.Stanford.EDU
Fri Feb 12 09:35:50 CET 2010


Hi Arthur/Karl --

Well, I checked (the most recent part of) my notes, and I think there's
a pretty good chance that what I say here is really what I've converged
on.  (It's possible I've forgotten something from earlier in my notes.)

 A> A cleaner, more thorough way of handling things would be for bibtex to
 A> understand what a TeX control sequence is,

 K> Or url.  I surmise it should only break lines at known good places, and
 K> there is no known-good place, it shouldn't break at all, overriding
 K> max_print_line.

My current plan is pretty much as Karl suggests here, where a "known
good place" is tentatively set to be defined as white space (but not
including the TeX tie character ~).  One issue I still need to check,
notes suggest, is whether Bill Gates's (evil) decision to allow spaces
in file names means that breaking at that space in a url (if it's
allowed in a url) will somehow mess things up.  I assume not, but as I
say, I want to check to make sure.

So can someone who really knows this stuff tell me whether breaking at
a whitespace character will ever be a real problem?  (I know it's
possible to redefine things in TeX to make all hell break loose, but
I'm not counting that as a real problem.)

[Historical note: The reason we (that is, I) got into this mess in the
first place -- possibly breaking a line in the middle of a TeX control
sequence -- was that I was assured by one of the TeX gurus (probably
this was a miscommunication) that it was always okay to insert a % and
then break the line at that point.]

 A> and to not break the line there.

 K> I think that would be much better than making max_print_line into an
 K> (semi-)dynamic parameter, or even a truly dynamic one.  Just because
 K> there is one 100-character control sequence or url in a bibliography
 K> doesn't mean I want every entry to be broken at column 100.

And my current plan is to leave max_print_line unchanged, with it's
default value still to be 79.

 A> Would such a change be considered meritorious or welcome?

 K> Meritorious.  Is it something you can work on?
 K>
 K> Background: I've talked to Oren quite a few times over the past several
 K> years about making a bibtex bug-fix release, with this incorrect
 K> line-breaking being the principal bug to fix.  Oren had hoped to do it,
 K> and still hopes to (not to mention BibTeX 1.0), but there is not even a
 K> STIX-like ETA at this point.  (Hope this is a fair summary, Oren.)

Yes, that's pretty accurate, except for an ETA for a bug-fix for just
this bug.  (Karl: I'll contact you offline within a day or two to
discuss the details.)

--Oren


More information about the tex-live mailing list