January 2023 Archives

This Week in PSC (095)

A busy meeting today, we talked about quite a few things:

  • Smartmatch deprecation continues. Some upstream PRs have been raised, awaiting CPAN releases
  • Refaliasing might be able to be deëxperimentalized if we add a warning on the currently-failing closure capture cases
  • RFC0013 highlights a deficiency in the overload.pm API shape. Perhaps an opt-in new calling convention is required to make it more flexible. Paul will write another post to the mailing list with more detail
  • Mithaldu’s objection to the suggestion to deprecate map EXPR, LIST suggests that maybe a more powerful debugger “run until next statement” command would be good
  • The interaction of List::Keywords + autovivification highlights the overall problem with highly-pluggable extensible systems - sometimes extensions conflict. We just have to keep this in mind and not have too high expectations that “everything will be fine if we load 20 different plugins”
  • That said, maybe there are some CPAN extensions that ought to be part of the core language - autovivification for example
  • We’ve run out of devel release volunteers now. We need some people to volunteer for 5.37.9, .10, .11, (maybe .12?). Also maybe 5.38.0

This Week in PSC (094)

PSC met today, all three of us attended.

We discussed:

  • HAARG’s map my $x RFC. Overall thoughts are good, with one or two minor questions we’ll add as comments.
  • Whether the additions to join() as part of RFC0013 should be gated by some sort of opt-in flag, so as to avoid surprises. Either a feature flag for the caller of join(), or a use overload import option.
  • Maybe renaming the RFC process itself (because of the ambiguity with IETF’s RFC) and improving the numbering system. More discussion needed.

This Week in PSC (093)

Paul, Philippe, and Rik had our regular call today to talk about perl5.

We put together some notes (at long last) for The Perl Foundation to use in their annual prospectus.

We talked about Math-Complex, and moving its issues into Perl5 GitHub. (A nice bit of work for anybody to do, but not groundshaking.)

We got back to the question of making Perl do SSL out of the box. Paul’s question was, “Can we bundle the CPAN dists for doing this and basically CPAN-install them at the end of normal install?” Audacious, but plausible!

We talked about the upcoming release of v5.38. Sure, it’s months away, but the various code freezes are approaching. What stuff needs work? One thing we did especially was review what experiments might be able to be deëxperimentalized. We talked about the refaliasing feature, which remains broken on closures, a total showstopper. Paul wonders if we can reduce the scope of how the feature can be used. For example: what if you must also use declared_refs?

This Week in PSC (092)

Having been off for two weeks, we spent a while just catching up with the state of the world. Not much of note to report this week.

  • We decided we should write up a list of “what’s expected in 5.38” to ensure we’re on track
  • We briefly chatted about :void, :scalar and :list as possible subroutine attributes for setting the return context

About Perl Steering Council