Re: [help-texinfo] alternate node name [was: Overloading function defs inside and across languages]

[Jean-Louis Leroy]

Hello,

I'm progressing in my documentation, and so far the really annoying
thing I've come across is the fact that node names also serve as node
ids and thus need to be unique (btw I've done a lot of work on
persistence and my experience says that deriving the id of a
persistent object from its state is always a bad idea).

I end up with needlessly long names that take no advantage of context,
e.g.:

@node Persistence in Relational Databases
@chapter Persistence in Relational Databases

The @dfn{relational tool} implements object persistence in relational
databases.

@menu
* Object Identity in a Relational Database::  
* Mapping Inheritance in a Relational Database::  
* Persisting Simple Types in a Relational Database::  
* Persisting References in a Relational Database::  
* Persisting Collections in a Relational Database::  
* Using Predicates with the Relational Tool::  
* Exercising the Relational Tool::  
@end menu

@Object Identity in a Relational Database

@node Mapping Inheritance in a Relational Database
@section Mapping Inheritance

etc etc etc

After a while I thought of adding a new parameter to @node, which
would be the 'display name' of the node. While tracing in makeinfo
I've come across cm_node() in node.c and if I've guessed right it
won't be easy at all.

I have the impression that makeinfo makes a single pass on the source
file, generating the HTML at the same time it parses. If that's the
case, I cannot access the @node command that defines the next node and
its new display-name argument.

Can anyone confirm or deny this hypothesis?

Also, while I still have great respect for the texinfo suite, my
current impression is that makeinfo kind of show its age. It's written
in C while it looks like other languages would be better suited (my
personal pick would be Perl). It does the lexing and parsing all by
itself instead of using tools like lex and yacc. It doesn't seem to
build a parse tree, in fact it looks a lot more like a simple macro
processor like cpp. I wonder when it was initially written, my guess
would be early seventies.

I hope that I'm not offending anyone here, I repeat that I'm very
happy that texinfo - including makeinfo - exists and I'm saying good
words about it around me. Also, these impressions come from a one-hour
gdb session so perhaps I'm completely wrong here. But if someone on
the list has hacking experience of makeinfo, I'd love to hear of him.
-- 
Jean-Louis Leroy
Sound Object Logic
http://www.soundobjectlogic.com