Back from holidays all three of us were in.
- We discussed the recently reported closure memory leak in 5.40, which existed in older perls and was fixed in a way that caused other breakage instead, and we agreed that if it isn’t fixable we need to address it by at least providing guidance on how to code around it
- We discussed the removal of apostrophe as a package separator, and agreed that it should likely be feature-guarded, but the verdict on the fallout is not yet in
- We continued the discussion on how to simultaneously encourage low ecosystem churn while helping users enjoy all the latest goodies
- We agreed on writing a few position paper articles to explain our thinking and the attitude we think a PSC should take
[P5P posting of this summary]
Just Aristotle and Philippe this time (Graham chipped in on IRC):
- we discussed the apostrophe situation: we will watch 5.41.3 break CPAN, and then evaluate the actual fallout. We like the idea of guarding this with a feature (which might need to be split in two, for the string interpolation case)
- we had a long discussion about backwards compatibility and
use VERSION.
Should "did you use VERSION?" become the new "did you use strict and warnings?"
[P5P posting of this summary]
Just Graham and Aristotle this time.
- Discussed some patches to include in a perl 5.38 or 5.40 point release.
- We will be watching the fallout from the removal of apostrophe as package separator.
- Discussed what could be done for lexical method calls.
[P5P posting of this summary]
Just Graham and Aristotle this time.
- Discussed tracking perl authorized releasers updates and storing historical data
- Dual-Life modules need some cleanup to get a more normal release process that can be done more regularly. Cleanup has been done on Storable.
- Time::HiRes by default doesn’t seem terrible, if possible
- X509 in core could make sense as part of a more complete interface to OpenSSL
- Discussion about how it would be possible to have support for Markdown or other new Pod extensions in a more backwards compatible way. Possibly a way to define alternate formats for
=begin/=end sections.
[P5P posting of this summary]
This week, the whole PSC was in attendance.
We merged HTTP-Tiny#6, and then discussed a number of topics:
- some improvements on the process for releasing blead-upstream dual-life modules are needed
- the instructions for releasing a new Perl are difficult to follow and partly redundant, the process takes longer than it should, and more of it could be automated
- the pumpkin permission list has grown over time, and probably needs to be trimmed
- more talk about Markdown documentation
- thoughts on how to strike a better balance with the verbosity/brevity of meeting minutes
