Re: index problem

[Eli Zaretskii]

> Date: Sat, 08 Sep 2001 22:58:09 +0200 (CEST)
> From: Werner LEMBERG <address@hidden>
> 
> I begin to understand.  My main concern was how the index looks like,
> not how to search an index entry.

To me, the index entries are the primary means of using a manual as a
reference.  In a well-indexed manual, I usually can find what I'm
looking for in a matter of seconds.

> BTW, I think I've found a bug (or at least a bad feature) in info 4.0.
> Considering groff again, there are two escapes called \a and \A.  Now
> suppose that I want to search index entries for \A: pressing `i', then
> `\A', and finally <tab>.  After entering the tab, `\A' has been
> converted to `\a'!  This is not correct IMHO.  If the user is
> specifically asking for an uppercase letter, info must not convert it
> back while searching index entries.

I will see what can be done here.  From my experience with hacking the
stand-alone Info, the case-conversion might be very hard to remove
without rewriting low-level search code.  But perhaps this case is an
exception...

Note that hitting TAB the second time shows both the upper-case \A and
lower-case \a entries, so Info will find both.

> It should work exactly as the isearch function in Emacs.

Well, it's hard to compare the stand-alone reader with Emacs in this
aspect, since Emacs doesn't yet support completion in `i', and so
pressing TAB at that spot has no effect.  But Emacs also finds both \A
and \a as well when you type "i \A RET", so the actual search is
case-insensitive in both readers.