lispref structure really confused (was Re: [Bug: xemacs-beta tip] tables of contents at wrong end manual)

[Samuel Bronson]

On Fri, Oct 8, 2010 at 1:55 AM, Stephen J. Turnbull <stephen at xemacs.org> wrote:
> Thanks for looking at this!
>
> Samuel Bronson writes:
>
>  > Hmm, it looks like there are worse problems, at least in lispref: the
>  > sectioning hierarchy doesn't match the node next/previous/up links
>  > doesn't match the node menu structure.

I'm having trouble finding this code. Perhaps there should be
directions for finding it in man/README?

> There is Lisp code in a comment in either the XEmacs User Guide or the
> Lispref (or maybe both) that is useful for maintaining both menus and
> links.
>
>  > I'd like to straighten this out a bit; any hints as to how to tell
>  > which order/structure(s) to keep?
>
> Use your good taste as far as order goes.  If as far as you can tell
> it doesn't matter, see if the helper code will decide for you (IIRC it
> either constructs menu from the node graph or vice versa) and if not,
> flip a coin.

Gotcha.

>  > One other thing I've noticed: usually, when I try "C-h C-f", it says
>  > that whatever I was looking for is missing from the index, but when
>  > the .texi file has a perfectly-good looking @defun (or similar) in it.
>  > Am I wrong in thinking that @defun is supposed to put the function
>  > described into the function index, or is something just mysteriously
>  > broken?
>
> AFAIK the primitive @def* directives handle this, so, yes, it should
> update the index automatically.
>
> Sounds like possibly Makefile breakage to me.  I don't use C-h C-f
> except by accident (I generally work with the manual source 'cause
> there's always something I want to change ;-), so I haven't paid much
> attention to that.

Too bad there's no way for the Info mode to take you to the source for
the text near point.

Hmm. Looking closer, I see that lispref has one index for everything.
Is it possible that this somehow confuses the `Info-index' code?