[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: texi2html build system

From: Reinhold Kainhofer
Subject: Re: texi2html build system
Date: Tue, 8 Apr 2008 12:07:16 +0200
User-agent: KMail/1.9.9

Am Dienstag, 8. April 2008 schrieb John Mandereau:
> On 2008/04/07, Reinhold Kainhofer wrote:
> > Ah, thanks, I didn't even notice that! The problem was that texi2html
> > added a space between <h2> and the section title, i.e.  "<h1> Chapter
> > blah </h1>", while , while makeinfo didn't use a space "<h1>Chapter
> > blah</h1>". I updated the regexp in that extracts the
> > section titles accordingly to allow spaces right after <h\d> and before
> > </h\1>... Docs are rebuilding on
> Thanks for this fix!  Is @translationof handling ready so I can convert
> translated docs?  

Well, for references within the same document, everything works fine (as 
texi2html keeps a hash of node<=>filename#ID assignments anyway). However, 
for cross-references from other documents (i.e. a link in the LM to the NR), 
there is no system yet in place to handle those. But that's not specific to 
@translationof, as it is already a problem with multiple nodes inside the 
same .html file. The @ref only points to the nodename, which might be inside 
a file that does not have anything to do with the nodename... We'll have to 
fix this problem in any case, so @translationof does not make things worse.

> I just pushed some nitpicks on make/doclang*.make,

Thanks, however, one thing about the --init parameter of texi2html: 
Unfortunately, it needs to come last in any call to texi2html, since the 
options to texi2html are interpreted sequentially and in our .init file we 
checks for several things defined by options. So, if  --split=section is put 
after --init=..., the .init script can't check whether we split by section or 
not at all, since the SPLIT variable will only be defined when the --split 
option is processed. Similarly, the --lang=.. should be before --init, 
otherwise we don't have the language available in the .init script.

That's the reason why I always put the $(TEXI2HTML_INIT) last. 

> and a (dirty) fix for 
> keeping <body> attributes in texi2html-generated pages.

Thanks. I simply ignored all attributes as a quick hack, as I didn't see an 
easy way to preserve the attributes using split.

> There are generally a lot of dangling xrefs in German and French docs,
> this is mainly my fault as I moved sections in all languages to match
> new GDP sectioning without checking xrefs.  I'll fix this as soon as
> possible (probably next week-end).

Yes, it might make sense to regularly run linklint or some other link checker 
on the whole documentation to check for dangling links (this might also help 
us find places where our texi2html code messes things up)...


Reinhold Kainhofer, Vienna University of Technology, Austria
email: address@hidden,
 * Financial and Actuarial Mathematics, TU Wien,
 * K Desktop Environment,, KOrganizer maintainer
 * Chorvereinigung "Jung-Wien",

reply via email to

[Prev in Thread] Current Thread [Next in Thread]