develooper Front page | perl.perl5.porters | Postings from March 2022

PSC #056 2022-02-25

Thread Next
Paul "LeoNerd" Evans
March 1, 2022 18:02
PSC #056 2022-02-25
Message ID:
Present: Neil, Paul, Rik

## trim vs trimmed

We discussed - again - the various naming arguments. While we're still
not entirely happy with "trim" it does seem overwhelmingly the popular
choice, and while two of us disagree with this, the role of the PSC is
to build consensus, rather than enforce our opinion. So for the sake of
making progress, we agree it can be called "trim".


It sounds feasible that we should be able to generate the AUTHORS file
from git commit history, provided we can ensure that extra names that
might be present don't get lost. We might need to keep a separate list
of those in an AUTHORS.EXTRA file or somesuch.

This will also require some updates to the release maintainer's guide
(as will perldelta; see below). "Yves has now volunteered to do this –

## BBC

The question was raised about whether we should add mention in
perldelta about a new release of DBIx::Class, which is needed because
the previous one was failing on latest 5.35.x. Historically there have
been "Incompatible Changes" sections added, so it seems appropriate we
could add one for this case.

## Criteria for Ending an Experiment

We've been continuing to refine down the policy surrounding
experiments, and exactly how we should handle the eventual
de-experimentation of particular features.

We discussed criteria for deeming an experiment as successful. We'll
iterate on our notes and then post them to p5p.

## Configure for taint

Neil is working on PRs to add a taint support question to Configure.
This will be at least two PRs: one for perl/metaconfig and one for
perl/perl5. Neil asked if it would be ok to split the perl5 changes
into two: a first one for Configure and perlsec.pod and one later for a
bunch of smaller pod updates. Rik and Paul were ok with that".

## perldelta naming

Dan Book proposed a change whereby releases would include the generic
`perldelta.pod` but also a copy named `perl5NNMdelta.pod` for that
specific release. We'd like to see this done for 5.36, if someone would
like to volunteer ... ?. We still think that change is good, but so far
nobody has made any progress doing it. This requires changes to the
Release Managers Guide, but also to various Porting scripts that are
invoked by release managers. Someone needs to (be) volunteer(ed) to do

## RFC Tracking

At this point in the call we had about 15 minutes remaining, which left
some time to review some more Pre-RFC suggestions.

 * printf format symbols for I32, U32:
     Not really big enough to justify an RFC to be honest. Just write it
     as a regular PR.

 * Check if running in debugger:
     Reject, as there are already mechanisms for that

 * Increase maximum line length to 100:
     Rik will follow up with an email

 * autorequire:
     Rejected: it's a dangerous feature, and if you really want this,
     there's at least one implementation on CPAN

 * builtin functions for num vs. strings:
     Paul will follow up with an email

 * UNIVERSAL::import:
     This RFC came in three parts.

     We like the first part (provide an empty 'UNIVERSAL::import' which
     warns of args), that should progress to an RFC.

     The other part - about automatic import ability - feels like much
     of an overlap of Ovid's `:export` attribute, and also steps on
     much prior art from CPAN. We still suggest someone needs to do a
     more in-depth investigation of existing solutions from CPAN in
     this space, and find a better core design to more fully cover it.
     (in particular on that note, it would be good to get any sort of
     lexical-import mechanism into core perl. We've needed it for a
     long time now)

Paul "LeoNerd" Evans      |  |

Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About