Re: cross references to manuals in different encoding fail

[Patrice Dumas]

On Sat, Aug 20, 2022 at 01:53:23PM +0100, Gavin Smith wrote:
> On Sat, Aug 20, 2022 at 02:24:29PM +0200, Patrice Dumas wrote:
> > Hello,
> > 
> > Cross reference with non ascii character fails when the other manual is
> > not in the same encoding.  I attach an example with an ISO-8859-1 manual
> > and an UTF-8 manual, both have the node
> >  @node chap Cross-ref thé
> > and an xref to the other manual, like
> >  @xref{chap Cross-ref thé,,,cross_ref_utf8}.
> > 
> > The cross reference with a non ascii character in the node name does not
> > work, in the latin1 manual, the error is:
> >  Cannot find node 'chap Cross-ref th\351'
> > In the UTF-8 manual, the error is:
> >  Cannot find node 'chap Cross-ref thé'
> 
> In my opinion this is not worth spending any time on.  In practice node
> names are nearly always all-ASCII.  

Not in non english languages.

> Nobody's ever reported a problem
> with inter-manual cross-references before, so I doubt it's a real problem.
> It shouldn't be surprising that Info files in different encodings can't
> refer to each other in this way.

I agree that it is not that much a real world example, collections of
non english manuals with cross-refs do not exist yet, if they start
existing they should be in UTF-8.

However, we should document it (it is actually already documented, but
not where users are likely to look).

> It seems wrong to go to a lot of effort to allow this use case, when
> it has never worked, and non-UTF-8 encodings should be on their way out
> by now.

Ok.  But this also means that we should document that the non UTF-8
encodings are not recommended anymore and are kind of legacy.

-- 
Pat