- Host: Google Hangouts
- Dates: Tuesday January 9th, 2018
- Times: 11:00pm–midnight Pacific Time
- Location: same Google Hangouts link as before
- Contact:
- Name: JF Bastien
- Email: [email protected]
None required if you've attended before. Email JF Bastien to sign up if it's your first time. The meeting is open to CG members only.
The meeting will be a Google Hangouts call.
- Opening, welcome and roll call
- Opening of the meeting
- Introduction of attendees
- Find volunteers for note taking (acting chair to volunteer)
- Adoption of the agenda
- Proposals and discussions
- Review of action items from prior meeting.
- Update on WebAssembly atomics
- Browsers are disabling SharedArrayBuffer due to Meltdown / Spectre, what does it mean for the WebAssembly equivalent? (@flagxor + @binji)
- Split some simpler subspecs out of the threads spec and let them land sooner? (@lars-t-hansen)
- Sign extension opcodes
- Mutable exportable globals
- Programmatic data initialization
- Discuss upcoming in-person meetings.
- Other items TBD
- Closure
None
None
- Alex Danilo
- Andreas Rossberg
- Anthony Laforge
- Arun Purushan
- Ben Smith
- Ben Titzer
- Brad Nelson
- Daniel Ehrenberg
- David Wrighton
- Deepti Gandluri
- Derek Schuff
- Eric Holk
- JF Bastien
- Keith Miller
- Kevin Cheung
- Lars Hansen
- Marco Trivellato
- Marcos Diaz
- Mark Nottingham
- Michael Holman
- Paolo Severini
- Pat Hickey
- Sean Westfall
- Sven Sauleau
- Thomas Nattestad
- Ulrik Sorber
Pat seconds.
- SIMD is in background work. Brad will report. Integer SIMD may be ready by April.
- Tail call: no updates. Can add as option to proposal. Some people oppose this. No discussion since in-person meeting.
- Exception handling: ongoing work.
- CSP: Google talked to Mozilla, trying to move forward. Some CSP experts on the hook. Apple folks like their approach.
- CSP and associated PR
- JS embedding tests: Dan started.
- WPT: nothing new.
- Browser limits: JSc updated their values to more closely match V8. Lars says Mozilla has a question about number of locals (did sum of params and locals). Ben Titzer will update proposal in spec repo.
Skip other items.
Update on WebAssembly atomics
Browsers are disabling SharedArrayBuffer due to Meltdown / Spectre, what does it mean for the WebAssembly equivalent? Slides
Brad gives summary of attacks
SAB can be used as a high precision timer (count on one thread, witness updates in another). There are other sources of timing in browsers. Brad presented a very early design that would re-enable SAB with much caveat (extra processes and such). Discussion about what other browsers think, it might be possible to just re-enable as-is.
“Our long-term plan is to make Spectre impossible even in the presence of high fidelity timing; further work is needed to re-enable this path.”
WebAssembly atomics are basically tied to SAB at this point. Waiting to figure out Spectre mitigations before talking about enabling.
Split some simpler subspecs out of the threads spec and let them land sooner?
- Sign extension opcodes
- Mutable exportable globals
- Programmatic data initialization
POLL pull each of these 3 items out into their own repo
unanimous
Discussion on feature testing of these features once split up. Threads could require all 3, and testing can just be pessimistic.
April would be a good time (Andreas would like early).
Fastly might host in San Francisco.
August timeline after that.
Google Munich would be able to host afterwards, in summer.
Igalia could host as well.
Zalando could also host, either Germany or Finland, maybe in November.
Luke pointed out the following after the CG call endded If WebAssembly has a presence at TPAC 2018 (October 22-26, in Lyon, France), we agreed last time we should reuse that as an in-person meeting opportunity.
Keep meetings bi-weekly, and rotate times.
Send invites in UTC.