Porters,
We had an abbreviated PSC call today, largely due to an unexpected delay.
We discussed offering split-up deprecation categories, so you can no warnings 'deprecated::.xyz'
and re-affirmed that we want to do this.
We talked about improving the backcompat of strict-vs-version behavior for use vX
where X is older than v5.36 and agreed we'd bring that back, but wanted to discuss more about other related changes to the use-vX code.
All three of us met, and were joined by Pete Krawczyk who first wanted to discuss some TPRF-related issues.
Following on from this we discussed:
Zefram's request to back out the strict-vs-VERSION changes. There's some merit here, he sent a PR to undo the whole thing. We feel there's two different issues that need looking at separately. Paul will explain some thoughts in more detail.
Yves's request to reöpen discussions about "deprecated::" warning categories. There may be benefit here too. Paul will write a followup response.
After missing last week, all three of us attended.
- Recapped the missing weeks from absences
- Smartmatch deprecation seems to be causing some BBC failures due to new warnings, but that’s to be expected; Philippe to submit patches upstream
- Refined the questions surrounding SSL-in-core; Paul to send a follow-up email
- RFC0013 needs some core changes to how overload works; best deferred for 5.39 now
- Agreed to rename “RFC” to “PPC” (“Proposed Perl Change”)
Paul and Philippe attended (Rik is still enjoying Australia).
We discussed:
- feature-class: there's a soon-to-be-merged PR to fix most memory leaks https://github.com/Perl/perl5/pull/20809
- n-at-a-time map/grep: HAARG should reply to any remaining open questions, so that we can likely move the RFC to the next step of the process ("Exploratory")
- SSL in core: it's still not in core, and there's an open PR to just dual-life the SSL modules, but that's unlikely to be the solution we're looking for. Also, we're still looking for a project manager for this project. Enquire Within.
- RFC0013: awaiting the ability to store more information in the interpreter about overload (for API version and flags)