Re: Perl 5.44.0 is now available

* Paul "LeoNerd" Evans <[email protected]> [2026-07-16 00:14]:
> This will also help us catch back up with our usual yearly release
> cycle, which as you'll note is now two months delayed this time.

    Jesse     5.12.0        2010-Apr-12
    Jesse     5.14.0        2011-May-14

    Ricardo   5.16.0        2012-May-20
    Ricardo   5.18.0        2013-May-18
    Ricardo   5.20.0        2014-May-27
    Ricardo   5.22.0        2015-Jun-01
    Ricardo   5.24.0        2016-May-09

    Sawyer X  5.26.0        2017-May-30
    Sawyer X  5.28.0        2018-Jun-22
    Sawyer X  5.30.0        2019-May-22
    Sawyer X  5.32.0        2020-Jun-20
    Sawyer X  5.34.0        2021-May-20

    ----------------------------------- PSC era
    Ricardo   5.36.0        2022-May-27
    Ricardo   5.38.0        2023-Jul-02
    Graham K  5.40.0        2024-Jun-09
    Philippe  5.42.0        2025-Jul-03
    Leon T    5.44.0        2026-Jul-15

So we slipped nearly two extra weeks from the previous most delayed
releases in this cycle, which is less dramatic than it seems but very
much not great at the same time.

More worrying (if that=E2=80=99s the right word; maybe unsatisfactory?) is=
 what
appears to be a trend when comparing era by era, particularly since the
PSC was established. Only the first PSC delivered a release in May, and
most releases by a PSC have slipped to July, whereas none in the pre-PSC
era ever did.

This year, we had all the LLM security findings to contend with, which
caused big changes to have to be merged during what was *supposed* to be
code freeze. A big delay resulted. Hardly surprising. I don=E2=80=99t know=
 what
lesson to take from this occurrence in particular: will we face this
kind of situation again =E2=80=93 a multitude of sufficiently-scary-to-unf=
reeze
issues reported all at once during freeze? Who knows. While the number
of issues found was out of the ordinary, it was not huge, and while some
of them were too serious to brush off, none were truly terrifying =E2=80=
=93 and
the advancements in LLMs up to this point have manifested in better SNR
(fewer findings with higher likelihood of being real issues) rather than
scarier bugs. So there are signs that this may not repeat again. However
my crystal ball is currently out of order.

But a higher-confidence conclusion is that the PSC should probably be
starting something, whatever that thing is, earlier in the cycle. I=E2=80=
=99m
not really sure what, but the release history seems to indicate there
is something.