Thanks to all who attended today's meeting. The next one will be on
Saturday, March 18 at 20:00 UTC.
What we did:
- Brian Heim and Nathan Ho worked on sclang LPC tests
- Nathan Ho also made (marginal) progress on UGen regression tests
- Patrick Dupuis investigated openbuildservice.org and HID on Fedora
- Plan for merging LPC test suite: get it to run, see if it catches
changes to sclang, quick review, merge
- openbuildservice.org, while good, is not a useful service for the SC
dev team since it creates a PPA/third party repository.
- Performance on MacBook Pro 2016
- Brian is pushing the Boost updates because it works towards fixing the
Windows helpfile problem
- Find people and get support for the SC_SCLANG flag:
https://github.com/supercollider/supercollider/pull/1939 - Mixed feelings on single quotes in variable names:
https://github.com/supercollider/supercollider/issues/2773 - Backing up SC:
https://github.com/supercollider/supercollider/issues/2776 - Rational numbers implementation: put off until 3.10?
- Push release date back to May 1
- sclang verbosity proposal is not a requirement for 3.9, but it's
- Brian and I share an opinion on the state of SC development right now:
we're overloaded with work, and the lists of issues and PRs often feel
like a huge wall of noise. We've been cutting down on multitasking and
trying to focus on what's important, and have felt increasingly
nonchalant about proposals for new features.
sc-dev mailing list
> - Brian and I share an opinion on the state of SC development right now:
> we're overloaded with work, and the lists of issues and PRs often feel
> like a huge wall of noise. We've been cutting down on multitasking and
> trying to focus on what's important, and have felt increasingly
> nonchalant about proposals for new features.
First, I'm sorry that I can't make the meetings. They are in the middle of
the night in China. That means I'm not helping with outstanding issues as
much as I perhaps should.
Second, I think it's totally reasonable to propose a (porous) moratorium on
new features until the backlog is a lot shorter, maybe even temporarily
closing issues and adding a "postponed" tag.
A roadmap would be a good idea, listing the top issues to handle in the
next month, and 3 months, and 6 months. Then stick to it, except maybe
critical crashing/hanging bugs (but I have to say, lately the SC core seems
a lot... a LOT... more stable and trustworthy than 2-3 years ago, so I
wouldn't expect a lot of those). I doubt there's any way to make the
backlog less frightening except to cut down the size of what we plan to
address in the near term.
"Do-ocracy" may not be the ideal strategy at present. It may be necessary
to start assigning issues. For myself, I'll probably contribute more if
there are specific issues on my plate, and less if it involves extra