develooper Front page | perl.perl5.porters | Postings from December 2021

PSC meeting notes, #45 #46 #47 #48 #49

From:
Ricardo Signes
Date:
December 20, 2021 00:55
Subject:
PSC meeting notes, #45 #46 #47 #48 #49
Message ID:
535abfaa-5539-4227-ac45-38ab4a633722@beta.fastmail.com
Porters,

I recently got an email saying, "Hey, is everybody okay over at the steering council? Awfully quiet." I apologize, we've definitely done a poor job of getting our minutes to the list in a timely way. I think all three of us have had our lives outside of Perl become somewhat more hectic as the year comes to an end, and while we're continuing to talk and have our calls, writing things up has gotten less attention than it needs.

Here, I am collecting the notes I have from the last few calls that happened.

PSC 45 - 2021-11-12

The three of us were here, along with Nicholas Clark. The great majority of this call was spent reviewing the status of the RFCs in flight and turning them into todos. Since this call, a number of things have come of that. Mostly I would call out the `kv` RFC, discussions on `defer` and `compile`, progress on booleans, a bunch of follow-up conversations about submissions or proposals, and some tweaks to our process docs.

One small non-RFC note: we discussed the proposal to add `-?` to perl as a synonym for `-h` and posted our approval on the pull request.

We also discussed the importance of updating code in core to use new features, especially those from new changes, where we could see how they'd feel to use in earnest. The first step, of course, is to make sure we finally adopt that weird new "use strict" feature. 🤔

PSC 46 - 2021-11-19

All three of us were present.

RFC follow-ups

We got together to follow up on actions the previous week's meeting, mostly in the form "send emails about RFCs" or "fix some known bugs in RFC-related code" or "fix our RFC-tracking spreadsheet." We each had one item left on the table, mostly "send an email" where we hadn't gotten to it. See also this delayed summary of meetings email!

Things we hadn't gotten done yet:

 * rjbs: follow up on source::encoding

 * paul: follow up on err

 * neil: follow up on Configure for disabling taint

PAUSE and CPAN

We talked about the way that CPAN's file signature checking wasn't performing up to its requirements, which led to reviving the still-ongoing discussion on p5p about shipping SSL support with core when possible. (I would love if somebody could make this happen and would buy them a nice dinner.)

Signatures

I said I would write an email saying "I really want us to get signatures out of experimental status" and then I did so, which has led to the current ongoing discussion about `@_` and DB::args and all that stuff.

Corinna

We were recently (at the time) asked, "What's up with Corinna? When is it coming?" My take is always, "I don't know, this isn't a p5p-led project!" But we made an action to email Ovid and ask him to join us next time so we could ask about it.

PSC xx - 2021-11-26

We didn't meet on November 26th. That was the day after Thanksgiving, and I certainly wasn't going to do anything but relax and eat leftovers.

PSC xx - 2021-12-03

We didn't meet in December 3rd. Honestly, I forgot about it until it was too late for me to make it. Paul said "Are we on?" and I said, "Oh crap," and we decided to punt until the next week, when I promised to be ready. (Spoiler: I made good on this.)

PSC 47 - 2021-12-10

Neil, Paul, and I met, and Ovid joined us.

The first and most import question was, "Ovid, how do *you* pronounce *Corinna*?" He said "cor-EE-na", which was good to know, because at least one member of PSC had been saying "cor-INN-uh". This sort of question really preoccupies me.

We talked about what next steps existed for Corinna to have a first deliverable. While we said (and I strongly maintain) that Corinna is not a p5p project, we were happy to offer our opinions, which largely agreed on "We'd love to see a minimal deliverable that people could start using, even if it is much less than half of what is promised."

We discussed the tensions at play, which I summarized as:

 * if we ship something incomplete, reaching completeness may require significant changes to designs shipped early

 * if we ship something early, people may decide it is stable and rely on it, especially if completion takes a long time to achieve (see: signatures)

Ovid said he would write up some possible next steps (which we did not expect to be a promise or even a plan). He has done so, but I haven't yet read them. (Thanks, Ovid!)

We intended for that discussion to take about 15 minutes, but it took closer to 55, so very little else was discussed.

PSC 48 - 2021-12-17

Neil, Paul, and I met on Zoom as usual.

First, Neil and I acknowledged that we needed to email Ovid back. Then I mentioned that we were definitely falling behind on sending out these summaries.

I asked about the progress on `@_` changes for signatures. I got a bit of a summary, but really the best summary may be "read the recent list posts." There are number of complicated trade-offs involved, and I'll keep my own opinions out of this summary.

We discussed `builtin` and its behavior as an experiment. My proposal was:

 * in v5.36.0, using anything in `builtin` is an experiment, because the whole thing is an experiment

 * in v5.38.0, only individual builtins will be experiments, like if we add `builtin::argc`

We discussed `finally {...}` and how it might work outside of the context of a try/catch block. I said — although I hadn't given it much thought — that maybe the compromise was to allow finally also to attach to any block that would be governed by redo/next/last. We (read Paul) don't have a clear plan here yet, other than "Paul will write an RFC."

Paul also believes that bools, RFC 0008, are totally done.

Meanwhile, Neil is on the hook for the Dec 20th release of Perl v5.35, and Rik promised to be around to help with any complications.

PSC 49 - 2022-01-07

Note that the date above is in the future! Our next scheduled meetings would be Christmas Eve and New Year's Eve, respectively, and we are exceedingly unlikely to have any Zoom calls.

-- 
rjbs


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About