Re: Fwd: [PATCH] different approach to --split html

[Eli Zaretskii]

> From: address@hidden
> Date: 09 Nov 2000 17:21:41 +0100
> 
> > Personally, splitting by node to me seems like overkill.  Splitting by
> > chapter has always been my personal preference.
> 
> Uh.  You can't do this very easily.

I have a working program that almost does that.  (``Almost'' because
it splits by the node, but the code which decides where to split can
be easily changed to do it by the chapter, or however else we decide
to do it.)

> We can only avoid a second pass
> by assuming that every reference is a node name and every node name
> can be mapped to a filename, before the node is being processed.

There's no requirement that everything be done in a single pass.  In
fact, makeinfo already does a second pass in some cases.

> Also, info files are (written and assumed to be) split by nodes.

No, they are split by size on node boundaries.

> (This is a reason why I have my doubts about the introduction of
> @anchor in 4.0.  It introduces references that are not nodes, not a
> trivial extension).

Could you please explain what problems does this introduce (except,
perhaps, requiring a second pass)?

IMHO, @anchor is a valuable feature, especially when you can produce
HTML.  @xref's to a node may land you very far from the place where
you want, and Web browsers don't even have the trick used by Info
browsers to search for the xref text after jumping to the location.