Re: [help-texinfo] is there a "Gits Guide" to debugging a texinfo file problem?

[Rick Jones]

Karl Berry wrote:
> Hi Rick,
>
>     I have the netperf manual in texinfo format:
>
> Cool.
>
>     http://www.netperf.org/svn/netperf2/doc/netperf.texi
>
> You mean http://www.netperf.org/svn/netperf2/trunk/doc/netperf.texi
> I assume.  (I've left /trunk out of more url's than I care to remember. :)

Grrr. Yes. :)

>     and somehow between revision 92 of that file, and revision 103 
>
> Can you send the file as it existed in those two revisions?

Might be a triffle large to send.  However, if I'm lucky, some of the 
tags may be good enough, just a second...

OK:

http://www.netperf.org/svn/netperf2/tags/netperf-2.4.3/doc/netperf.texi 
matches the overall revision 92 version.  There is no tagged version 
that equals revision 103 but:

http://www.netperf.org/svn/netperf2/tags/netperf-2.4.4/doc/netperf.texi

is beyond that version and still has the problem.

>     I have managed to do something that causes it to "drop" "Using
>     Netperf to Measure Aggregate performance" when I C-u C-c C-u m in
>     emacs (21 or 22).  I'm presently at a loss as to what I may have
>     done and was wondering if there were some handy debugging guide?
>
> I'm afraid not.  The only tactic that comes to mind is the usual --
> simplify the file as much as it existed.

I was afraid of that.  I've never really understood how the texinfo 
syntax was supposed to work, I was only ever just bluffing my way 
through it :(

> Looking at the file, one thing that comes to mind is to drop the
> pointers off all the @node lines (so they contain just the name of the
> node).  I am doubtful that it will help, but at least it is a
> simplification (and makes searches in the source less painful).

The theory being that the existing previous next and up "links" may be 
confusing things?  Sorry if that is an obvious question.

thanks,

rick

> Another possibility if we can't get it is to ask address@hidden; the
> Texinfo support in Elisp is maintained there, these days.  Lots more
> Emacs hackers hang out there than here.  (I'm sure they will appreciate
> seeing the two versions too.)
>
> Best,
> Karl