2022-05-03 [14:20:49.0080] Richard Gibson: published npmjs.com/parseint and added some fixes to a few others, and now i'm down to 54 failures. (attached, if you're curious) [14:21:14.0142] * Richard Gibson: published npmjs.com/parseint and added some fixes to a few others, and now i'm down to 54 failures. (attached, if you're curious) [14:22:16.0143] the vast majority are Date ones; one pair is "parseInt.prototype isn't undefined" because that's not shimmable; another pair is unavoidable AggregateError constructor hijinx; and then a few scattered others [14:32:00.0702] Awesome! I haven't looked at the other failures, but for parseInt, the `asNonConstructible` of https://matrix.to/#/!dGIsaFIukRWLdpurvh:matrix.org/$fvGRXxCTo23zwsh5a2s7KyK_ZHFuYv0msf3o2YmKEDE?via=matrix.org also takes care of making the prototype undefined in environments supporting Function.prototype.bind. [14:35:38.0926] oh right, i haven't tried that one yet [14:38:17.0187] hm, that doesn't seem to be working for me [14:39:32.0059] ``` function p() {} Object.prototype.hasOwnProperty.call(p, 'prototype') // true Object.prototype.hasOwnProperty.call(Function.bind.call(Function.call, p), 'prototype') // still true in node 0.4 [14:43:11.0360] hmm, I wonder at what Node version that changed [14:45:37.0954] how do you handle observable behavior that _becomes_ shimmable in general? 2022-05-05 [09:08:15.0967] feature-detect it, and shim it whenever possible [09:08:38.0648] i only tried in node 18, maybe it works in older engines? i'll check [09:08:53.0073] psa: i'm making an executive decision to cancel the maintainers meeting during TC39 in june 2022-05-06 [07:19:57.0937] > <@ljharb:matrix.org> i only tried in node 18, maybe it works in older engines? i'll check I see bound `call`s having no own "prototype" property for any version of Node.js on RunKit (0.10.48 through 17.9.0): https://runkit.com/rstpflkbkgqetqlavy/bound-call-own-prototype [08:37:27.0369] hmm, maybe my testing was wrong, i'll look thanks [08:41:10.0654] ok interesting. however that means that the result has the name "bound call" unless function names are configurable, and also `new` of the result doesn't throw [08:41:28.0796] so it seems like it'd maybe pass the letter of the test but not the spirit [10:38:06.0530] that's why I asked about observable behavior that becomes shimmable... but `bind(nativeCall, fn)` _does_ cover both non-constructibility (i.e., exceptions from `new`) and absence of an own "prototype" property, and does not preclude name preservation: https://runkit.com/rstpflkbkgqetqlavy/bound-call-constructibility it really seems to be the best option in environments that lack Proxy and arrow functions. 2022-05-08 [13:54:12.0598] > <@gibson042:matrix.org> that's why I asked about observable behavior that becomes shimmable... but `bind(nativeCall, fn)` _does_ cover both non-constructibility (i.e., exceptions from `new`) and absence of an own "prototype" property, and does not preclude name preservation: https://runkit.com/rstpflkbkgqetqlavy/bound-call-constructibility > > it really seems to be the best option in environments that lack Proxy and arrow functions. ok so thanks, this does work as expected in any env where function names are configurable. that precludes IE and node < 3, for example. there's also a few engines where _bound_ functions, specifically, don't have names. so in any of these i'd probably have to just skip the attempt at not having a prototype 2022-05-09 [09:59:53.0288] FWIW, https://runkit.com/rstpflkbkgqetqlavy/bound-call-constructibility shows [bound] function name configurability in node v0.10 and v0.12. There are almost certainly _other_ environments where even this approach will fail, but that is itself detectable so "shim it whenever possible" per https://matrix.to/#/!dGIsaFIukRWLdpurvh:matrix.org/$pxkgqqJrnk2SUTxQ2xugHyqrzjOlDXaQFC7dTuyIrCQ?via=matrix.org remains available. [16:52:38.0844] still having some trouble; see the "wip" commit on https://github.com/es-shims/parseInt/actions 2022-05-10 [09:42:39.0532] I'm not sure how to investigate those tests... can you reproduce the failure locally? 2022-05-11 [17:32:48.0322] yes, it reproduces locally. You check it out, run npm install, and npm run tests-only 2022-05-12 [11:36:55.0921] Hi everyone! At this week's meeting, we came up with a plan to try to get through the last two big chunks of in-person discussion (priority of constituencies and values-based decision making) over the next two weeks, so that we can have the new frameworks ready for June plenary. [11:37:45.0157] However, we need everyone to be able to be around for those calls, so we'd like to confirm we can come or reschedule. ptomato has mentioned he cannot make the 26th, so I will send a doodle for earlier that week. [11:38:23.0393] **In the meantime, can you please mark here if you can or cannot make the usual time on May 19.** [11:39:03.0361] (I can make it, I just wanted to set up the emoji vote.) 2022-05-16 [10:03:50.0889] @room Doodle for times this week: https://doodle.com/meeting/participate/id/dwp3Vpze [10:10:07.0159] Doodle times for next week: https://doodle.com/meeting/participate/id/e1wyWERb 2022-05-17 [03:57:43.0880] Rick Waldron, jugglinmike, ljharb please fill out the doodles. 2022-05-18 [09:00:39.0126] done, sorry for the delay 2022-05-19 [10:44:17.0872] Hi everyone, I'm not going to be able to make it to today's meeting. 2022-05-26 [05:16:24.0554] regrets for today from me! (I think I mentioned it last week already) [11:04:21.0964] so far i'm the only one in the jitsi [11:05:14.0629] seems like today might be cancelled? [11:08:28.0252] k, richard, mike, and i were there, we had no agenda, and bailed early. see yall next week 2022-05-27 [00:59:40.0421] Unfortunately, the meeting time is too late for me. If I could make it, I would like to know people's thoughts on https://github.com/tc39/test262/issues/3524, suggested by Rick Waldron