Re: info-lookup-symbol

[Eli Zaretskii <eliz@gnu.org>]

> From: Rehceb Rotkiv <rehceb@no.spam.plz>
> Date: Sun, 08 Apr 2007 18:07:07 -0500
> 
> when I use info-lookup-symbol, I sometimes get to the wrong place in the 
> info file. For example, when I use info-lookup-symbol on the keyword 
> "break" in a Python source code file, it takes me to the index item
> 
> * break <1>: try statement.
> 
> instead of
> 
> * break: break statement.

Confirmed.

> probably because (strangely) the first appears before the latter in the 
> index of the info file.

That's the reason, yes.  The ones with <N> appear before the one
without because this is how disambiguation of identical index entries
is implemented in Texinfo; I will ask the Texinfo maintainer to change
that.

> How can I avoid this?

You can edit the index of the Python manual by hand, putting the main
index entry first in the Index node.  I don't have any better ideas at
this time.

But the _real_ fix is to complain to the authors of the Python manual:
they shouldn't be having several index entries named identically in
the first place.  They should instead qualify all but the single entry
(the one that points to the "break statement" node) with the
specialized aspects of "break" described in the other places.  For
example, the "break <1>" entry could be renamed to something like
"break, in `try' forms".

> Is it, for example, 
> possible, to get a summary of all index items that contain "break" 
> instead of directly jumping to the first one?

You can switch to the *info* buffer and type "i break TAB".

> Also, how does Emacs know 
> in which index file to look / how can I customize it?

This is set up either in info-look.el or in the language-specific mode
package.  In this case, see python-after-info-look in python.el.