groff 1.24.0 plans? (was: most widespread complaints about groff 1.23.0

[G. Branden Robinson]

Hi Colin,

At 2024-07-04T08:56:01+0100, Colin Watson wrote:
> What are the plans for a 1.24 release?

I muttered some thoughts a few months ago.

https://lists.gnu.org/archive/html/groff/2024-02/msg00054.html

> In general I think it's bad to drift into a habit of allowing too much
> time and too many commits to pass between releases;

That's true, and I still like the idea of getting 1.24 out this calendar
year, 1.25 in 2025, and so forth...

> it makes it more difficult to work out what went wrong when users
> report regressions,

Yes, Git bisections of the formatter itself get tedious even with a
modest growth rate of O(log n). ;-)

> and it invites a "just one more fix" mindset which tends to be
> never-ending until the complaints pile up high enough.

I may suffer from this mindset regardless of time elapsed since release.

:-O

There is a 1.24 release goals Savannah ticket.

https://savannah.gnu.org/bugs/?65099

I should probably file a few tickets to add to it:

1.  [troff] next-generation alignment/adjustment control

    See https://lists.gnu.org/archive/html/groff/2024-06/msg00053.html
    and follow-ups.

2a. Drop cp1047.tmac, de-document it, and NEWS its withdrawal.

    See:

    https://lists.gnu.org/archive/html/groff/2023-03/msg00113.html
    https://lists.gnu.org/archive/html/groff/2023-04/msg00000.html
    https://lists.gnu.org/archive/html/groff/2024-05/msg00028.html

2b. [troff] emit special diagnostic on first occurrence of input
    character code > 127

    This would tell people that EBCDIC support is gone, that they
    need to load one of the specialized macro files for character
    encoding support, and that those will be going away too in groff
    1.25 in favor of support for UTF-8 input, so they need to look into
    preconv(1).

    In my opinion we need something like this since the semantics of
    input code points > 127 will change from groff 1.23 to 1.25.
    People/distributors skipping over 1.24 won't get warning,
    unfortunately.

    Unless they do something crazy like read the release notes.

None of these are difficult or time-consuming, and the first is already
partially done, including the hardest part (separating alignment from
adjustment data inside the formatter).  I've been waiting to see if
any protests were lodged.

I'll shoot Bertrand an email and see if he's amenable to conducting
another release.

Regards,
Branden

signature.asc
Description: PGP signature