> with the last fixes in texinfo.tex, the output looks much better
> now, the @verbatim indentation is fixed,
Yes.
> the URL spacing is sometimes better, sometimes worse than with the
> old texinfo.tex version, IMHO.
This I still have to investigate. Will do that in the next few days.
> The issue with the additional space in front of manually sorted
> index entries is fixed for most entries,
Indeed, but there remain problems for non-letter cases like '!' or
'<'. Will report this to Gavin soon, too.
> the problem persists for entries like
>
> @funindex @sortas[fooA} foo
> @funindex @sortas{fooB} \foo
>
> however.
It's the same issue, I guess. BTW, I've just noticed that '\bracket'
appears twice in Appendix E of the NR. Reason is that someone added
@funindex \bracket
to `expressive.itely` without a proper '@sortas' tag (right now this
is an educated guess, but I'm rather sure about that).
To avoid that we should probably add a script that does the following.
(a) Scan the documentation to check whether there is both
@funindex foo
and
@funindex \foo
at the same time. Emit warnings for such entries, telling the
user that '@sortas' must be used.
(b) Scans the documentation for index entries that have '@sortas'
and warn if similar entries are found without '@sortas'.
> I tried to understand how this whole macro expansion magic works,
> reading the 'TeXbook' back and forth, but this is currently beyond
> my depth.
:-)
> Do you think this one can be regarded as a regression in
> texinfo.tex, too?
Yes.
> I'm asking, because I was not sure, if the way we use the @sortas
> macro is intended. (Defining our own macro @funindex that uses
> @sortas and then putting in another @sortas between @funindex and
> the argument...)
Well, I don't think that our '@funindex' macro is the culprit. I
guess it is rather a general problem with @sortas, and hopefully Gavin
will find a work-around (or a real fix).
Werner
