PSC #019 2021-05-06
Present: Rik, Nick, Neil
We discussed the Perl 7 version bump - “reports of my death have been
greatly exaggerated”:
The plan remains that there will be a Perl 7 bump, but not immediately after
5.34.0 is released. We want to make it “good value” - to have clearly
communicable improvements. We don’t think that we can deliver on that in 12
months. Hence
* The intent is that we continue to do timeboxed stable releases annually
* When we hit “all the 7 features” then it’s 7 - i.e. most of the changes
for 7 may well be released first in 5.x releases.
* 7 won’t be infinitely far away - if it turns out that a feature will take
too long, then that’s for 7.x.0, or even 8
* for now the focus is on the obvious low-hanging fruit that can be ready for
a 5.36.0
* developing the plan for 7 remains on our list, and some of the 5.36 work
may result in things deemed "for 7".
5.34.0:
https://github.com/Perl/perl5/issues/18739 - Nick takes this
https://github.com/Perl/perl5/issues/18764 - Nick triages this
We’d like to have fixes for these in good time for Sawyer to roll RC2
5.35.x:
We had a brief discussion about dual life modules.
* Neil had a module with a change in blead, and had missed this for some
time, until nudged by Todd doing 5.33.9 (thanks!). Can we handle this
better?
* There are dual life modules whose canonical repository needs to be a
subdirectory within blead. It’s often hard to make a CPAN release of these
- too many seem to have weird one-off requirements. Can we simplify/better
automate any of this? Rik and Nick both hoped Neil would do this, but he’s
not as daft as he looks.
Rik to start working through the list of experiments, to see how to bring
each to an end. He’d like to start with what seem to be the simplest/least
controversial, but some of these are entangled with experiments which will
need more thought, so picking where to start is actually hard.
Improving perl5-porters:
Neil to start the public discussion on this [which has already happened]. We
chatted briefly about the IRC channel #p5p, but didn’t have any
conclusions. We had an amusing digression about Usenet newsreaders:
Nick: Gnus
Neil: rn
Rik: slrn
We think that Neil wins that one :-)
Other:
Nick is continuing research into RFCs, PEPs and similar from other language
communities.
We will investigate how to “smoke CPAN” to get some quantitative idea of the
impact of proposed changes (features and deprecations) - Nick and Rik are
both interested in this, and will talk to others who have attempted it
previously, or might have useful insights.
Nick is getting better at not talking too much - for the first time Neil did
not have to signal a “time out”.
Present: Rik, Nick, Neil
We discussed the Perl 7 version bump - “reports of my death have been
greatly exaggerated”:
The plan remains that there will be a Perl 7 bump, but not immediately after
5.34.0 is released. We want to make it “good value” - to have clearly
communicable improvements. We don’t think that we can deliver on that in 12
months. Hence
* The intent is that we continue to do timeboxed stable releases annually
* When we hit “all the 7 features” then it’s 7 - i.e. most of the changes
for 7 may well be released first in 5.x releases.
* 7 won’t be infinitely far away - if it turns out that a feature will take
too long, then that’s for 7.x.0, or even 8
* for now the focus is on the obvious low-hanging fruit that can be ready for
a 5.36.0
* developing the plan for 7 remains on our list, and some of the 5.36 work
may result in things deemed "for 7".
5.34.0:
https://github.com/Perl/perl5/issues/18739 - Nick takes this
https://github.com/Perl/perl5/issues/18764 - Nick triages this
We’d like to have fixes for these in good time for Sawyer to roll RC2
5.35.x:
We had a brief discussion about dual life modules.
* Neil had a module with a change in blead, and had missed this for some
time, until nudged by Todd doing 5.33.9 (thanks!). Can we handle this
better?
* There are dual life modules whose canonical repository needs to be a
subdirectory within blead. It’s often hard to make a CPAN release of these
- too many seem to have weird one-off requirements. Can we simplify/better
automate any of this? Rik and Nick both hoped Neil would do this, but he’s
not as daft as he looks.
Rik to start working through the list of experiments, to see how to bring
each to an end. He’d like to start with what seem to be the simplest/least
controversial, but some of these are entangled with experiments which will
need more thought, so picking where to start is actually hard.
Improving perl5-porters:
Neil to start the public discussion on this [which has already happened]. We
chatted briefly about the IRC channel #p5p, but didn’t have any
conclusions. We had an amusing digression about Usenet newsreaders:
Nick: Gnus
Neil: rn
Rik: slrn
We think that Neil wins that one :-)
Other:
Nick is continuing research into RFCs, PEPs and similar from other language
communities.
We will investigate how to “smoke CPAN” to get some quantitative idea of the
impact of proposed changes (features and deprecations) - Nick and Rik are
both interested in this, and will talk to others who have attempted it
previously, or might have useful insights.
Nick is getting better at not talking too much - for the first time Neil did
not have to signal a “time out”.