Add agenda for WG to advance js-string-builtins to phase 5 #1844
Conversation
|
Let's advance other proposals to phase 5 too. AFAIK there's no requirement that this be done by the champion (I assume that's why @eqrion singled out string builtins). @lukewagner @rossberg @conrad-watt @tlively @aheejin @titzer @sbc100 @fgmccabe I don't know of any reason why EH, Memory64 and JSPI shouldn't advance to phase 5. IIRC there was some spec work to be done for threads... although it looks like there's progress there, so maybe in a month that will be ready too? |
|
It looks like we're a little inconsistent about whether we merge proposals before or after phase 5. AFAICT, EH and Memory64 have been merged to the wasm-3.0 branch, but JSPI and string builtins have not. @rossberg, can you confirm? I agree with @eqrion that the documentation says that we go to phase 5 and then merge. |
|
Yes, my understanding is also that we go to phase 5 and then merge. Strictly speaking, nothing on the wasm-3.0 branch is actually merged, since we are publishing from main. I'd like to get that fixed soon too, since I've seen cases of people confused by the large discrepancy between what we are publishing and what we are shipping (and it's quite inconvenient to link to specs for anything not on main). But I think that is a separate issue. |
|
@dschuff Okay, so if we're okay with merging phase 4 proposals to wasm-3.0, that should unblock my PR to the spec repo. I still would like to advance to phase 5 just to close it out completely though. I agree that the other phase 4 proposals should probably go to phase 5 too. I could roll them into my agenda item if you'd like. Also I'm not entirely sure what even needs to be presented or if this is a rubber stamp sort of thing. |
|
Merging JS string builtins into the wasm-3.0 branch SGTM. I would not merge 3.0 into main before all proposals are moved to phase 5 by the WG. But that should be a quick affair. Many people have asked, and I'd like to get 3.0 merged ASAP. The draft in wasm-3.0 has a few non-trivial todos though, working on which would be wasted effort if we do the switch to SpecTec. The SpecTec version in turn is almost done, except for a few remaining todos about prose generation (Wasm-DSL/spectec#141). @f52985 |
|
@rossberg, do you have an ETA for the SpecTec version? If it's going to be ~months, then getting the current wasm-3.0 branch merged sooner is probably worth it because the current situation is so rough. Let me know if there's anything I can help with, too. |
|
@f52985, what do you think, can we give a rough ETA? |
Yeah, I’ve started reviewing the TODOs and resolving a few of them. |
Per the discussion in WebAssembly#1844, we should move the rest of these proposals to phase 5 while we're at it.
Per the discussion in #1844, we should move the rest of these proposals to phase 5 while we're at it.
|
@f52985, any update on the SpecTec TODOs? |
We are currently working on them. However, new issues have been emerging during the refactoring process, and some are more complex than initially expected, which is causing some delays. We aim to resolve all of them by the end of July. Progress is being tracked in the following GitHub issue: Wasm-DSL/spectec#141 |
5 participants
No description provided.