On Wed, Sep 30, 2026 at 08:02:29PM +0100, Gavin Smith wrote:
> > To me this is what we decided to do, but we could also have decided to have
> > different rules for insert_nodes_for_sectioning_commands and shadow
> > targets, this consistency is not a necessity.
>
> I don't really understand this point, as there should be rules for what is
> a valid Texinfo input file regardless of the options used for texi2any.
> For example, whether @xref{foo} is valid or not depends on whether "foo"
> is a target.
There should be Texinfo language rules, but there could also be
differences among output formats, and ways to do differently from
the rules. We have this situation for the validity of menus, for
instance, we have rules, but also ways to avoid the rules for a
non-hierachical manual.
--
Pat
