[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: cross manual references in html manuals
From: |
Eli Zaretskii |
Subject: |
Re: cross manual references in html manuals |
Date: |
Sat, 24 May 2003 11:48:22 +0300 |
> Date: Tue, 20 May 2003 12:53:13 +0200
> From: Dumas Patrice <address@hidden>
>
> > and '.' is left as is. I think '.' should also be considered
> > special, and also the file names should not be related to html,
> >
> > Sorry, I don't understand either of these points. How is . special?
> > How are the file names related to html?
>
> . is special in files names, as it is traditionnally used to separate the
> file name from the extension.
IMHO, this is only so in archaic filesystems. In modern filesystems,
`.' is just another character, with the last `.' serving as some
indication of the file's type/format/purpose. (Texinfo doesn't break
this special meaning of the last period in the way it produces HTML.)
> > And, I think this is achievable, because Eli implemented it for
> > makeinfo. We just generate an anchor for each node, as in (putting the
> > node name as the text of the anchor here is not functional, but anyway
> > ...):
> > Node: <a name="$foo">$foo</a>,
> > ...
> > Node: <a name="%25foo">%foo</a>,
> >
> > Then the references to the two nodes become -foo.html#$foo and
> > -foo.html#%25foo, respectively. Do you see a problem with this?
>
> Yes, it cannot be valid xhtml. In xhtml, only [A-Za-z0-9-_] are
> acceptable as text for the name= or id= attributes.
Is XHTML relevant to Texinfo? (Sorry, I don't know what XHTML is.)
If XHTML is relevant and should be supported, I think it will be very
hard to reconcile it with split-HTML mode and cros-manual references.
It's not hard to write code that mangles a node name into something
that only uses [A-Za-z0-9-_], but I don't see how such mangling could
be consistent across manuals. That is, the hard problem is to ensure
that a different instance of makeinfo that is processing a different
manual will come up with exactly the same mangled name. Right now, we
use a node name with a few simple and deterministic changes, so the
references we produce are consistent.
> Using 8-bit characters in node names raises another issue, which is
> how to translate these characters in file names ?
Same as now I suppose: convert them to `-'.