| 06:46 | <ljharb> | also it's a US holiday |
| 15:16 | <bakkot> | Michael Ficarra: do you want to review my extra commits to https://github.com/tc39/ecmarkup/pull/666` |
| 15:16 | <bakkot> | not a request, you are welcome to not |
| 15:23 | <Michael Ficarra> | LGTM |
| 16:57 | <bakkot> | so, this PR requires abstract method names to be globally unique. right now, certain environment records have a GetThisBinding, and others do not, which cannot be specified with this PR |
| 16:57 | <bakkot> | I guess I will just make all of them have it and the ones which currently don't will simply assert that they are not called? |
| 17:01 | <Michael Ficarra> | WFM |
| 17:50 | <bakkot> | there's a few cases where we're invoking an abstract method on a type which is statically knowable |
| 17:50 | <bakkot> | it would be nice if there were a way to make the generated link go to the correct concrete definition without just writing it out manually (which breaks the "references" machinery) but I can't think of one offhand so I'm not going to worry about it for now |
| 18:10 | <Michael Ficarra> | explicitly emu-xref the abstract method with an attribute that describes the concrete implementer? |
| 18:12 | <bakkot> | I think that breaks the "references" thing? but actually maybe not, I'll have to check |
| 18:13 | <Michael Ficarra> | no I think explicit emu-xrefs show up in references |
| 19:22 | <bakkot> | should we auto-generate ids for the trs in the table? alternative is to hand-write them |
| 19:22 | <bakkot> | I want links to go to the relevant row rather than the whole table |
| 19:41 | <Michael Ficarra> | hand write them |
| 19:46 | <bakkot> | most AOs end up being like sec-methodname but sec is mostly for clauses; thoughts on abstract-methodname here? with "abstract-" as a prefix for all of them |
| 19:46 | <bakkot> | gonna go with that for now |
| 20:07 | <Michael Ficarra> | You know I don't care about IDs. They just have to have some kind of vague relation to the thing so that if you see two spec URLs with fragments and you know the two places they're supposed to go, you can usually tell which one is associated with which place without clicking. That is the only requirement. |
| 21:10 | <bakkot> | Michael Ficarra PTAL https://github.com/tc39/ecma262/pull/3745 |
| 22:13 | <Michael Ficarra> | @bakkot PRs LGTM |