On Tue, Jul 28, 2026 at 9:48 PM G. Branden Robinson <[email protected]> wrote: > While it's too much fuss and bother for groff 1.25.0, longer term I > think I'd prefer to work around that difficulty in "test-groff.in" and > any test scripts and Makefile recipes requiring a "full" $PATH; in > exchange we'd get a more deterministic build of our documentary > artifacts--which the reproducible builds movement may appreciate--and > greater sensitivity to a botched generation procedure for said > artifacts. > > What do you think?
I've argued before for the usefulness of test-groff as a general-purpose groff testing script, and use it that way frequently myself. So I'm not keen on a change that makes it less useful in that realm. However, I already lost that argument some six years ago (http://savannah.gnu.org/bugs/?57630), and your proposal seems in line with the scope of test-groff as envisioned in that ticket, so I can find no new grounds upon which to object.
