Re: WNPP pages not using slices properly

[Denis Barbier <barbier@imacs.polytechnique.fr>]

On Mon, Oct 16, 2000 at 05:14:20PM +0200, Josip Rodin wrote:
> On Mon, Oct 16, 2000 at 04:58:46PM +0200, Denis Barbier wrote:
[...]
> > If you need more complicated stuff, a trick is to dynamically build slice
> > specifications (i.e. they will be processed by slice but not wml), e.g.
> > 
> >  %!slice -o UNDEFuADOPTu$(LANG):being_adopted.html -o UNDEFuPACKu$(LANG):being_packaged.html ... 
> > 
> > where $(LANG) (or any other variable) is replaced during pass 1.
> 
> I'd use CUR_ISO_LANG but that's `en', not `EN', and the manual page says
> that the sliceterm can consist only of uppercase letters. :/

In the sliceterm grammar, 'n' is for slice intersection, and 'u' for
union. I think this is why lowercase letters are forbidden within slice
names.  If you really need the second solution and are not afraid, try

  <define-tag slice-lang>{:[[tr/a-z/A-Z/]]$(CUR_ISO_LANG):}</define-tag>
  %!slice -o UNDEFuADOPTu<slice-lang>:being_adopted.html
  %!slice -o UNDEFuPACKu<slice-lang>:being_packaged.html
  ...

I do not have webwml here at work, so i may miss an obvious solution,
will tell you later.

Denis