| 18:39 | <keith_miller> | robpalme: Hello are you around? |
| 18:50 | <robpalme> | hello Keith! |
| 18:51 | <keith_miller> | robpalme: Dan mentioned that for top level await you wanted microtasks to be delayed so it'd be easier to inline modules into the requester(s)? |
| 18:52 | <keith_miller> | I wasn't sure I fully followed that. Maybe you could explain it to me? From what I understood it didn't seem like it would work if you wanted to inline modules with more than depth 2? |
| 18:53 | <keith_miller> | Since once you finish the second dependency you have to turn the microtask queue according to the current spec |
| 18:53 | <keith_miller> | e.g. if a -> b -> c there needs to be a microtask turn between b and a |
| 18:54 | <keith_miller> | If you want I can forward/CC you to an email thread between Dan (+ others) and me, which maybe will help provide context |
| 19:01 | <robpalme> | please forward it on |
| 19:02 | <robpalme> | the original request (promoted by guy bedford) was that synchronous module graphs should remain fully inlineable (i.e. concatenated into one run-to-completion) to preserve what Rollup does today |
| 19:04 | <keith_miller> | Oh! Yeah, that makes sense. I also think it won't be web compat to change that |
| 19:04 | <keith_miller> | I thought this was about async subgraphs |
| 19:05 | <devsnek> | you want to insert a delay into TLA? |
| 19:05 | <robpalme> | right - guy had a later request that i agreed with on TLA - need to dig it up |
| 19:05 | <keith_miller> | I don't have your email can you DM it to me rob? |
| 19:06 | <keith_miller> | devsnek: It's to drain the microtask queue between each module in an async subgraph. |
| 19:06 | <keith_miller> | right now it drains effectively randomly |
| 19:06 | <robpalme> | https://github.com/tc39/proposal-top-level-await/issues/158#issue-796269657 <-- i think this is the issue |
| 19:07 | <keith_miller> | Ah, ok. Yeah I read that issue |