[l2h] Another take on the 'no images' problem

[Jean-Pierre Chretien]

>>From: Ross Moore <ross at ics.mq.edu.au>
>>Date: Thu, 8 Jun 2006 17:06:47 +1000
>>To: Robin Fairbairns <Robin.Fairbairns at cl.cam.ac.uk>
>>X-Brightmail-Tracker: AAAAAQAAAAQ=
>>X-Whitelist: TRUE
>>Cc: latex2html at tug.org
>>Subject: Re: [l2h] Another take on the 'no images' problem
>>
>>Hi Robin, and others.
>>
>>On 06/06/2006, at 6:32 PM, Robin Fairbairns wrote:
>>
>>>   One way to fix the problem is to have
>>>       \pdfoutput=0
>>>   at the top of your LaTeX jobs, but there should
>>>   be an easier way to configure the same result.
>>>
>>> no, after loading the package that breaks things.  the problem is  
>>> that some
>>> packages (including html.sty, last i looked) use a broken means of  
>>> detecting
>>> that they're using pdflatex, and on modern distributions this  
>>> brokenness means
>>> that it *always* decides it's using pdflatex, regardless of how  
>>> latex was
>>> called.
>>
>>Yes, it's true that  html.sty  does have a slight problem in this  
>>regard,
>>but that *should not* be relevant to the generation of images by
>>a LaTeX2HTML job.
>>
>>The  images.tex  file that is created by LaTeX2HTML should not include
>>lines that cause either  html.sty  or  hyperref.sty  to be loaded.
>>
>>However, that being said, it *is* possible that a user's own package
>>or \input source might be requiring some coding that has a side-effect
>>of causing such a package to be loaded as part of the processing
>>of  images.tex .
>>
>>The *easy* way to avoid this is to identify lines of coding that lead
>>to this effect, but which are irrelevant to any images that need to
>>be created. Then simply surround these lines by "comments" as follows:
>>
>>   %begin{latexonly}
>>   ...
>>   ...
>>   %end{latexonly}
>>
>>This causes the intervening ... lines to be ignored by LaTeX2HTML,
>>while having no effect at all when LaTeX processes the same job.
>>
>>
>>
>>>
>>> i've mentioned this before here.  is there a way to distribute a  
>>> "working"
>>> html.sty, or do we have to wait until someone (presumably poor old  
>>> ross)
>>> has time to produce a new distribution?
>>
>>Yes;  html.sty  needs an update.
>>To be fully compatible with *verrrry* old LaTeX software, the test for
>>whether pdfTeX is being used should be something like:

Sorry if this is off-thread, but referring to my message of March 2006
http://www.tug.org/pipermail/latex2html/2006-March/003150.html
I wonder if this html.sty update can solve the problem, which appears with an upgrade
of a recent latex distribution (texlive-2004->texlive-2005)  ?
In fact I suspect not, as I get the same error with pdflatex or latex compilation,
but if html.sty should be updated, this other problem could be considered.

The compilation error I get is this:
! Extra \else.
\H at old@ssect ...t {#1}{#6}\@@par }\endgroup \else 
                                                  \def \@svsechd {#5{\hskip ...
l.1 \begin{thebibliography}
                           {xx}
                           
As I indicated, the ifacsam.tex and ifacsam.bib files are on CTAN. Is this reproducible elsewhere ?


Kind regards

-- 
Jean-Pierre