Re: Duplicate index entries in NR

[Trevor Daniels]

Phil Holmes wrote November 19, 2011 2:34 PM

> From: "Trevor Daniels" <address@hidden>
>
> > Phil, you wrote Saturday, November 19, 2011 11:44 AM
> >
> > > While looking for information in the NR I came across a 
> > > duplicate index entry.  A bit of copy-and-paste and text 
> > > manipulation and I find that there are 65 - I've got the full 
> > > list but a simple example is acciaccatura.  Is there any reason 
> > > for this?
> >
> > I don't understand what you mean.  I see two entries
> > for acciaccatura in the index pointing to two different
> > places where acciaccatura is mentioned in the NR.  This
> > seems right to me.  Do you have another example?
>
> If you look at accent, the two entries there are shown against a 
> single index entry, with 2 pages in the list of references.  I 
> believe that's how the acciaccatura should appear as well - 
> otherwise there's a tendency to miss the fact that there are two 
> entries and only follow the one, missing important information.

The @index and @funindex macros generate calls to
different texinfo macros.  From memory the @index
macro calls @kindex and @funindex calls both
@kindex and @findex.  The two types of index entries
are printed with different typefaces, so maybe this
is why they are kept separate in the pdf.

Actually the indices in the html manuals are in a far
worse state.  Entries are not only unmerged but
every @seealso generates a spurious unhelpful entry.
See, for example, ambitus.  Indexing has been discussed
several types in the past (search devel for index
funindex for a selection), but always deferred until
(a) the manuals were in a better shape and (b) the
syntax had stabilised.  Maybe one day ...

Trevor