Status and path to Phase 4

[yuri91]

I would like to see this proposal going forward, since it's a prerequisite for the branch hinting proposal.

We recently solved the main issue of missing tests (#15) by adding to the interpreter a way to do (optional) extended checks on custom sections/annotations (#17).

There are still 3 open issues, but I am not sure if they are active concerns or not.

Is there anything that still needs to be done to go forward? If so, I am happy to help.

[rossberg]

I think the only real open question is how this would interact with future repeated sections (#11). But I think we at least have a possible plan, as stated on the issue.

Other than that, this proposal was mainly held back by various meta considerations (#5), most of which were actually outside the scope of this proposal. I'd be happy to move forward.

[yuri91]

Thanks for the recap!
Nice to know that there are no real blockers.

[rossberg]

Well, not from my perspective anyway. If it was for me, this could have been merged years ago. ;)

[tlively]

The new text parser in Binaryen (not quite enabled by default) supports parsing annotations in general, although not the name annotations or the syntax expressing arbitrary custom sections. Assuming we will remove the name annotations as discussed in #21, then we should be able to declare the phase 4 toolchain requirement fulfilled once Binaryen enables the new parser by default and adds support for the arbitrary custom section syntax.

Are we planning to waive the Web engine implementation requirement?

[tlively]

Bump. @rossberg, I'd like to get this voted to phase 4 at the June CG meeting if possible. What else needs to happen before that? Do we need to have a CG discussion to resolve the question about the @name annotations?

[rossberg]

Sounds good! Yes, the only thing we need to resolve is whether to keep @name as the textual representation for the name section. We can try to resolve that beforehand or at the June meeting.

[tlively]

Great, I've proposed an agenda item for the next CG meeting so that I can find out sooner rather than later if I'm going to need to implement name annotations in Binaryen before the June meeting. WebAssembly/meetings#1554

[rossberg]

I'm afraid I won't be able to attend the next meeting, since I will be at a meeting in Japan.

[tlively]

How about the next one on May 21?

[rossberg]

That works.