| 15:47 | <nicolo-ribaudo> | Editorial question: if I were to write We use that notation in https://tc39.es/ecma262/#sec-innermoduleevaluation step 2.b. In that case it's guaranteed to always contain a throw completion, but seeing the syntax made me wonder how it would work. |
| 15:51 | <ljharb> | hm, it would for a variable, but I’m not sure the macro actually works in that case - if it does I’d say yes, it mutates the slot |
| 15:52 | <nicolo-ribaudo> | Maybe we should replace Return ? _foo_.[[Value]] with Return Completion(_foo_.[[Value]]) to avoid ambiguity in case somebody copy-pastes it in a case where a normal completion is possible |
| 16:07 | <Michael Ficarra> | I actually think we forgot to add a ? expansion for aliases (which we use in many places) |
| 16:07 | <Michael Ficarra> | ReturnIfAbrupt handles them fine, but no the ? shorthand |
| 16:09 | <Michael Ficarra> | assuming we made the ? shorthand work, the ReturnIfAbrupt expansion should work this way |
| 16:11 | <Michael Ficarra> | this conversation reminds me, I really want to get https://github.com/tc39/ecma262/pull/1573 in |
| 17:46 | <bakkot> | I do not think of ? as modifying anything, although I agree the current wording is ambiguous |
| 17:46 | <bakkot> | I actually want to get rid of ReturnIfAbrupt and just define ? and ! explicitly |
| 17:46 | <bakkot> | which would make this more clear |
| 17:46 | <Michael Ficarra> | yes, I will update #1573 to do that |
| 17:47 | <Michael Ficarra> | ... whenever I get time |
| 17:48 | <bakkot> | just added a comment to https://github.com/tc39/ecma262/pull/1573 which I think does away with the ambiguity |
| 17:52 | <Michael Ficarra> | I mean, it's just a single form (let) |
| 17:52 | <Michael Ficarra> | it's missing Perform ? ..., Return ? ..., F(? ...), etc |
| 17:53 | <Michael Ficarra> | also, what happened to tmp5? 😖 |
| 17:59 | <bakkot> | hey it's still valid, no one said the synthetic aliases had to have any particular names |
| 18:00 | <bakkot> | and the idea is that the other cases are handled by the last sentence:
i.e., |
| 18:00 | <bakkot> | though I guess it doesn't quite manage to actually say that |
| 18:01 | <bakkot> | I think that sentence does correctly handle F(? ...) though |
| 18:01 | <bakkot> | I should just say "in another context" rather than "in a more complex expression" |
| 19:54 | <TabAtkins> | rip tmp5 :( |
| 20:27 | <Michael Ficarra> | interview with @canadahonk on Porffor: https://www.youtube.com/watch?v=OSYSXq7y0Ew |