17:38
<nicolo-ribaudo>
During the Module Harmony intersection semantics discussion, those of you who are championing/authoring proposals please feel free to jump in the discussion as if you were co-presenters when there are discussion about your proposal :)
17:39
<nicolo-ribaudo>
( Luca Casonato Kris Kowal guybedford caridy )
18:38
<littledan>
Kris Kowal: I don't understand your comment; I thought this presentation is the epic
18:38
<littledan>
but yeah the presentation would be reworked and written down in other places maybe if that'd help
18:39
<Kris Kowal>
I’ll clarify that my comment is about organizing repositories.
18:42
<shu>
here's the one-sentence question form of my concern: how hard will it be for an implementer who's not going to any of the module harmony calls to design and implement one of these features?
18:44
<littledan>
here's the one-sentence question form of my concern: how hard will it be for an implementer who's not going to any of the module harmony calls to design and implement one of these features?
Do you have any concrete source of concern that it'd be harder for modules than other proposals?
18:44
<littledan>
the idea is to document things
18:44
<nicolo-ribaudo>
here's the one-sentence question form of my concern: how hard will it be for an implementer who's not going to any of the module harmony calls to design and implement one of these features?

I can speak for module expresions/declarations.
Right now, those two proposals "pretend" to be developed in isolation and don't call out anywhere in their docs their relation with the other proposals. This is something I can fix, to prevent the implementation problems you are concerned about.

Other proposals (i.e. the various compartments layers) do a better job with regard to this

18:44
<littledan>
(as Kris said)
18:45
<shu>
Do you have any concrete source of concern that it'd be harder for modules than other proposals?
yes, because half of it is in blink/gecko/WK, not in the JS engine
18:49
<nicolo-ribaudo>
Do you know how much this division follows the 262-HTML division? I'm sure things got mixed up with the hooks refactor we landed some months ago, but did the previous spec layering match the implementation layering?
18:52
<shu>
nicolo-ribaudo: sorry, i don't currently, would need to do research
18:56
<Kris Kowal>
I think the word we’re looking for is foreshadowing. Each individual proposal should have notes on mechanisms that foreshadow later proposals.
19:46
<nicolo-ribaudo>
import source from from; 👀
19:46
<nicolo-ribaudo>
(in practice no one would use from there and the ambiguity is solved by a two-tokens-lookahead)
19:57
<ljharb>
fwiw, wrt alternatives, i would not be happy with increasing the asymmetry between static and dynamic, compared to the current spec (reacting to some of the suggested alternatives)
19:59
<Kris Kowal>
I am in favor of parity in general between static and dynamic import.
19:59
<littledan>
Are these significantly different in terms of their asymmetry? I haven't heard anyone advocate for import.source foo from bar but certainly it'd be easier to parse
19:59
<ljharb>
(lol also, import.phase source foo from 'path' is an alternative that would work for import.phase() because dynamic import is syntax not a function reference)
19:59
<Kris Kowal>
The only distinction should ever be static discoverability.
19:59
<ljharb>
or import.source i suppose, sure