| 15:56 | <shu> | bakkot: so i finally got around to implementing the removing-await-from-yield* thing, and it showed some performance regressions, which is very strange. that should totally not be the case, right? is there any interaction with promise combinators that we missed? |
| 16:00 | <Ashley Claymore> | Purely curious, were these performance regressions of running “normal” sites/apps. Or synthetic? |
| 16:02 | <bakkot> | shu: no interaction with promise combinators I can think of, no. |
| 16:03 | <bakkot> | I can't see the performance regression page linked on the v8 bug; can you mark it as visible or send me a screenshot or something? |
| 17:42 | <Jack Works> | Purely curious, were these performance regressions of running “normal” sites/apps. Or synthetic? |
| 18:13 | <shu> | Ashley Claymore: synthetic afaik |
| 18:13 | <shu> | bakkot: that infra i believe is google only, i'll take a look later in the week |
| 18:44 | <jschoi> | I use async generators all the time to implement async-iterator transforms like mapping, filtering, and grouping. Though I use JavaScript for offline data processing, not for web apps. |
| 18:54 | <bakkot> | if you have access to github's new codesearch it is a great way to answer this question |
| 18:54 | <bakkot> | https://cs.github.com/?scopeName=All+repos&scope=&q=language%3Ajs+%22async+function*%22 |
| 18:55 | <bakkot> | finds e.g. https://github.com/prettier/prettier/blob/f38111fec6c35b513370832a84bdac8b5663763d/src/cli/expand-patterns.js#L13 https://github.com/PipedreamHQ/pipedream/blob/66ad30207aa5205021c37871008dd5b6d469a0a9/scripts/findDuplicateKeys.js#L20 |
| 18:55 | <bakkot> | second one has a yield*, even |