At 2025-10-24T13:28:11-0400, Peter Schaffter wrote:
> On Fri, Oct 24, 2025, Bjarni Ingi Gislason wrote:
> >   It is referred to with '\*[...]' instead of '\\*[...]'.
> > 
> >   Found with "test-groff -ww ... doc/automake.mom".
> 
> mom files should be processed with pdfmom(1), not test-groff.

That's a more categorical statement than I was expecting.  (Though you
did say only "should", not "must".)

I had thought that pdfmom was only _required_ if there were forward
references in the document.  That would be the same reason that the old
DWB troff mm(1) command existed,[1] and why we have mmroff(1) in
groff.[2]

> pdfmom does not emit an error/warning on \*[PDFBOOKMARK.NAME] (i.e.
> where the backslash is not escaped).  Moreover, the string is only
> called during "Lists of..." collection.  automake.mom has no "Lists
> of..."

I think another, possibly more basic, problem with Bjarni's command line
is that he used `-ww`, and mom(7) was not written in expectation of that
warning level (as we've discussed before, I think in Savannah tickets).

As far as I know, mom(7) in our master branch is `-wall`-clean,[3] and
so if no diagnostic appears when using that option (or any subset of the
categories it selects, including the empty set), even with "test-groff"
instead of pdfmom(1), then there is no problem.

My advice to Bjarni would thus be to stop using `-ww` with `-mom`,
regardless of formatter program or wrapper thereof.

Regards,
Branden

[1] https://github.com/n-t-roff/DWB3.3/blob/master/misc/mm/mm.sh
[2] https://man7.org/linux/man-pages/man1/mmroff.1.html
[3] https://savannah.gnu.org/bugs/?66122

Attachment: signature.asc
Description: PGP signature

Reply via email to