Back to the full roster.

• We talked again about Perl 42. We think it may already be too late for it in this cycle, so we want to make a thorough case for it by releasing a side tarball for the end of this cycle, and if no technical blocker is found, actually do the jump to 44 in the next cycle.
• We listed out next steps for the next point release.
• Aristotle summarized the suggested simplification of our plan for TLS in core, but a fuller discussion next week is necessary.
• PPC 0014 (English names) has no implementer.
• There was a branch for PPC 0021 (Optional chaining) but the implementer hit a wall and got stuck.
• Regarding the dual PPCs 0030 (equ) and 0031 (eq:u), we lean towards the former. Options would be cleaner if we had a big and regular set of them across operators, but neither is the case, and they constitute an extra concept.
• Many thanks to Dave Cross for providing an initial implementation of a PPC index page. We decided to merge it as is, then iterate on the PPC process afterwards, instead of doing it the other way around.

[P5P posting of this summary]

Just Graham and Philippe this time.

• We talked a bit about PPC 0025 (Perl version), but didn’t say anything new
• Regarding PPC 0030 (new operators) vs PPC 0031 (flags on operators), we continue to think that flags on operators bring more confusion than value (PSC #168). Perl is an operator-rich language, so adding new operators when needed sounds right.
• As for the discussion on implicit stringification of references, we think that adding more runtime effects to strict doesn’t really fit. The issue would likely be better addressed with a new warning (and is very similar to the uninitialized warning when dealing with undef).

[P5P posting of this summary]

Three of us again. Aristotle had limited time, Philippe and Graham stayed longer.

• Steve Hay released the RCs for for 5.38.3 and 5.40.1. Thank you!
• Philippe rebased his Perl 42 branch on top of blead after the release of v5.41.7. Things are on track to build a 41.8 testing tarball alongside the 5.41.8 release and we will keep this dual track going for now. We agreed that if we decide to go through with the 42 plan, we must have at least three 41.x-only releases, to keep with the feature freeze policy.
• We also considered the behavior of use VERSION when given a non-existent version from the gap between the last v5.41.x and 42. We think it should be an error, to continue helping with typos (in line with the current helpful error message when requesting e.g. 5.10 instead of 5.010).
• We talked about forbidding non-overloaded references stringification. Everyone seems in agreement that the feature is good to have, but the way to enable it is less clear for now: it could be a stricture, warning, feature flag, or pragma.
• We briefly touched on feedback on our preliminary plan for TLS in core, suggesting that an even simpler approach may be possible. We will pick this back up in a future call.

[P5P posting of this summary]

Just Aristotle and Graham for our first meeting of the new year. Not much progress since the last one due to Christmas, New Year’s, sickness, and other personal circumstances. We discussed our framing of the version bump, the timeline for a decision, and the fact that constraints push us toward a dummy .0 third version component as the simplest way forward.

[P5P posting of this summary]