At 2025-01-27T16:32:03-0600, G. Branden Robinson wrote: > At 2025-01-27T17:46:14+0000, Colin Watson wrote: > > It should be a matter of removing .gitmodules and instead setting > > GNULIB_REVISION in bootstrap.conf to refer to the desired Gnulib > > commit ID. You can then change that with ordinary commits as > > needed, rather than trying to remember the submodule commands. > > This is exciting. I'll keep in mind for early in the groff 1.25 > cycle. Dave's found yet another thing I managed to regress about 6 > months ago (again, not something anyone's likely to exercise or notice > in man pages), so I'm not eager to kick groff's build system in any > way.
I fixed that one (Savannah #66723) but Dave found another, one that alarmed Peter as well. (That is the hyphenation language globality item discussed this past week on this list.) > Trixie is drumming her fingernails and popping her bubble gum at me > impatiently, while Chim-Chim hoots and leaps about. I feel it necessary to disclose that while getting 1.24 ready on the timetable Colin required for the Debian Trixie release freeze (about 8 March) looked like a tight squeeze a couple of weeks ago, I now have no confidence that it can be met with anything more than a perfunctory rc1, and if we needed to roll an rc2 we'd be simply sunk. Judging by past experience, an RC needs to percolate for 2-3 weeks, and after that there's little evidence that any further useful feedback about it comes in. Absolutely everything would have to go right--and my maintainerly powers to upload things to GNU servers are as yet untested--and since the 27th something else has gone wrong, this time Savannah #66758. Another change from last October has proven to have untoward and unintended consequences. (This time it's a change to the mm package's debugging interface of all things. With any luck it's just a dumb syntax error.) This is somewhat frustrating but it's better to find these issues before making an RC, let alone a final release. Better now than later. Thanks go to Dave and Deri for giving the master branch thrashings that exposed issues that existing automated tests did not. So while I'll continue to work toward a release candidate and final release as soon as is feasible, I think that's going to be on the other side of the Debian toolchain/central infrastructure freeze deadline that Colin asked me to meet if I wanted to see 1.24 in Trixie. It seems that a curse afflicts the synchronization of the Debian and groff release schedules. (Maybe the curse bears my name.) It's flattering to think of groff in the same light as gcc and binutils, though. :) Regards, Branden
signature.asc
Description: PGP signature
