[20710] in Athena Bugs
Re: latex2html in the infoagents locker
daemon@ATHENA.MIT.EDU (t. belton)
Thu Sep 12 13:21:04 2002
Date: Thu, 12 Sep 2002 13:21:02 -0400 (EDT)
From: "t. belton" <tbelton@MIT.EDU>
To: Stephen Gildea <gildea@MIT.EDU>
cc: Nick White <nickw@MIT.EDU>, <bug-infoagents@MIT.EDU>
In-Reply-To: <200209112355.TAA17573@nerd-xing.mit.edu>
Message-ID: <Pine.GSO.4.33L.0209121312030.20198-100000@iphigenia.mit.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Ho boy. I apologize for continuing this cc list but I think I want this
recorded in the bug-infoagents archive - I may need it later. And I figure
the original recipients might want to see this.
OK. So. The configuration script in the new latex2html is broken. Not the
'configure' sh script itself, which is the same old configure everyone
uses. After that parses the switches, it passes them to config.pl, which
was apparently home-grown by the latex2html folks. This is the code has a
bug.
It's set to look in the right place now for the graphics utils, but its
existence check is still failing.
configure output:
checking for pnmcrop... no
Use of uninitialized value in concatenation (.) or string at ./config/config.pl line 1268.
Checking version of
Warning: Will not be able to generate images due to above failure.
See the line that says "Checking version of" nothing? I added that as a
diagnostic. The nothing is supposed to be the full path of the 'pnmcrop'
that they are checking the -version of. The Perl warning above it is
because they're trying to check something they haven't put anything into
yet.
Actually, I don't guess it's a bug - but the subroutine that puts the path
into that variable doesn't like our universe for some reason. So to us
it's a bug.
This will take a little while to work through. I also need to go poke
around the latex2html site and see whether there are any notes about this.
It's never simple.
