Meeting 2015 11 30

Lars Bergstrom edited this page Nov 30, 2015 · 1 revision

Agenda items

  • Brainstorm ideas for the workweek (larsberg)
    • DESTROY ALL INTERMITTENTS; Browser.html polish; alpha Android release hacking; GH issue backlog triage; rebase and merge abandoned PRs; ...
  • NCSU deadlines: Dec 4 is pencils-down, Dec 14th is expected window for addressing review comments. (jdm)
  • crate squatting (jack) https://github.com/servo/rust-layers/issues/217#issuecomment-160684294
  • PR queue is >80 again, almost 50% reviewable (ie. S-awaiting-review) (jdm)

Attending

  • jack, simonsapin, larsberg, ajeffrey, mbrubeck, manish, fewsxcv, edunham, pcwalton

Regrets: Ms2ger, jdm

Last week

Brainstorm ideas for the workweek

  • jack: I'm fond of destroying the intermittents!
  • larsberg: I suspect there are only three or so
  • pcwalton: ipc-channel improved failure messages?
  • larsberg: I think that plus the error reporting one would be good. I think ipc-channel gets a lot of blame due to chain-panic thread failures.
  • jack: Now have a version of the runtime that launches Servo, via ms2ger!
  • simonsapin: What's the runtime?
  • jack: browser.html requires something that isn't a browser to start from. It's just a mozjs shell. The first bit of it is support for the CEF API in JS, so that the JS code can launch a browser window. Can also be used to build apps that aren't browsers (e.g., commandline apps). We'll see where it goes. We have a first version and now are waiting for feedback from paul. Most of the blockers in terms of Servo layout bugs are either fixed or in-queue.
  • jack: Does anybody have other ideas/opinions?
  • manish: nox is saying the issue backlog is causing people to wonder how we feel about it.
  • jack: I think until we destroy intermittents, the PR log is going to stay long. After the intermittents, we can help with rebases, etc.
  • jack: Maybe land sandboxing?
  • pcwalton: There's some work needed on OSX, but it's landed.
  • jack: Could probably test all platforms in mozlando. I'd like to come up with a 2016 testing plan, too. What do we want to do? Need jgraham there (webdriver support, etc.). Also mochitests and bholley. See what we need to do to get servo well-tested. Fact that there are no cookie tests is a bit scary! Breaking cookies and having nothing catch it in WPT is scary.
  • manish: I think in the last few months started doing work on areas that don't have WPT tests. We're hitting a limit in some areas where we are missing tests.
  • jack: So far, we've been relying on WPT. Unless we start budgeting time to write new WPT tests, we won't remember to do it. In the initial 2016 CTO office roadmap planning, I did propose having a full-time person to work on testing stuff... btw, 2016 plans should be finalized after the workweek.

NCSU deadlines

  • jack: I assume jdm will want help reviewing them. Dec. 4 is code complete; Dec. 14 is the landing time. May want to do some of these on the workweek.
  • jdm: Please keep in mind that reviews/answers for NCSU work should be performed ASAP to maximize chance of usable, valuable contributions from students.
  • jack: jdm may pick reviewers and push them around.
  • simonsapin: Do we wait for students to address the issues in the review?
  • jack: Yes, but they have to be addressed by Dec. 14.
  • manishearth: Is merging a requirement for the course?
  • jack: I think not (it wasn't in the past); they had to get a PR opened. Extra credit before December 4th. We've had trouble in the past with not helping the NCSU students land their PRs during that last push window.

Crate squatting

  • jack: We lost the skia crate! Somebody took it. So now we're stuck as servo-skia. Simon registered all the stuff that we were not already using.
  • simonsapin: I sent some e-mails to the person who took the skia crate name, but they aren't spending any time on it and want PNACL support, which we don't have.
  • jack: Compile skia with PNACL? I talked to them back when that team still existed, and they weren't planning to support something like skia, even before that team got killed.
  • pcwalton: Maybe it's about pepperGL instead?
  • simonsapin: The person said there's a bunch of llvm, rustc, etc. stuff to do.
  • jack: I'll try to figure out from my google contacts if PNACL is still under any active development, particularly towards supporting skia.
  • jack: There's a list of crates not owned by us that are either being squatted or that have some other implementation. Maybe open a metabug and start tracking these down? Kinda bad to have servo-foo for every foo in the servo GH org. Also, if you create a new crate, SQUAT IMMEDIATELY!
  • simonsapin: Also don't forget to add the github team as an owner so that other people can publish updates. Run cargo owner -a github:servo:cargo-publish

PR queue is > 80 again

  • jack: Lots are in a reviewable state; please get on these! If you have time, pick up stuff that doesn't have a review. Might be hard before we have intermittents solved.
Clone this wiki locally
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.