On Sun, Sep 06, 2026 at 04:24:35PM +0100, Gavin Smith wrote:
> On Sat, Aug 29, 2026 at 02:30:56PM -0600, Karl Berry wrote:
> > Aside: maybe the unbalanced parens should actually be an error? And
> > detected by makeinfo? Since otherwise the "continued" text becomes part
> > of the definition. E.g., source like this does not come out right in TeX:
> >
> > @defmac AM_PROG_VALAC (@ovar{minimum-version}, @ovar{action-if-found},
> > @ovar{action-if-not-found})
> >
> > It has to be continued with a lone "@" at the end of the @def line.
>
> It doesn't come out right with texi2any either. However, it's likely
> to cause more problems for users to make it a hard error when it processed
> successfully before.
>
> texi2any doesn't appear to warn about unbalanced parentheses. I have no
> objection to it doing so, but somebody would have to do the work to implement
> that. I suspect it would not be completely straightforward to implement
> due to text being contained within various elements in the parse tree for
> the def line.
I think that it is better to let the user decide if unbalanced
parentheses are right or not. To me this is too dependent on the
language, parentheses could be used for something else than enclosing
arguments, or could be part of a constant shown as argument...
If this is really relevant, I think that it would better be outside of
texi2any, the SWIG interface would be better suited for that kind of task
in my opinion.
--
Pat
