00:14:50  * srl295quit (Quit: Connection closed for inactivity)
00:17:58  * keithamusquit (Ping timeout: 245 seconds)
00:20:10  * keithamusjoined
01:51:58  * AtumTquit (Quit: AtumT)
02:29:34  * Aquaziquit (Quit: Connection closed for inactivity)
03:56:38  * jmdyckquit (Remote host closed the connection)
08:28:43  * jwaldenquit (Quit: ChatZilla 0.9.92-rdmsoft [XULRunner 35.0.1/20150122214805])
10:13:10  * AtumTjoined
13:14:00  * jmdyckjoined
15:13:04  * Aquazijoined
15:22:27  * cloudshujoined
16:47:17  <devsnek>rwaldron: is the intention of Atomics.notify that Atomins.wake is gone
16:47:19  <devsnek>or is it just an alias
16:47:40  <rwaldron>Atomics.notify replaces Atomics.wake
16:47:59  <devsnek>and so the plot thickens
16:48:08  <rwaldron>How engines go about doing that is up to the implementers
16:48:15  <devsnek>i guess v8 isn't removing Atomics.wake
16:48:25  <rwaldron>Seems sane
16:48:35  <rwaldron>I wouldn't either
16:48:39  <rwaldron>not immediately anyway
16:51:21  <Bakkot>Hope they're adding a deprecation warning at least?
16:51:26  <Bakkot>Elsewise we end up with both forever
16:51:52  <devsnek>i'm just trying to figure out how to track this in node with v8 lol
16:52:07  <devsnek>this seems to be something of a first
16:55:00  <gsathya>I plan to add use counters for this, but maybe we'll end up with both forever
16:55:40  <devsnek>right now i think my best path is to switch the alias around so that Atomics.notify is just Atomics.notify and then Atomics.wake is an alias of notify that also emits a warning
17:43:50  <Domenic>I would put 60% odds on "both forever"
18:16:35  * keithamusquit (Quit: The Lounge - https://thelounge.github.io)
18:17:14  * keithamusjoined
18:37:23  * keithamusquit (Quit: The Lounge - https://thelounge.github.io)
18:39:19  * keithamusjoined
19:00:25  <Bakkot>gsathya: there's only like ten people in the world using Atomics; can we just talk to them directly?
19:02:07  * akirosequit (Ping timeout: 240 seconds)
19:02:47  <rkirsling>😂
19:05:29  <ljharb>gsathya: atomics were removed completely, and that was deemed web compatible - why would you need the alias at all?
19:05:34  * akirosejoined
19:06:10  <ljharb>iow it's impossible that anyone is relying on code using atomics.wake on the web, right, since all the browsers unshipped it?
19:07:05  <gsathya>because feature detection for atomics, doesn't necessarily involve feature detection for atomics.wake
19:07:53  <ljharb>so the premise is that there's sites that work without atomics, feature detect their presence, and then blindly use .notify when present?
19:08:08  <ljharb>i guess i see how that's plausible but i find it very unlikely
19:08:20  <ljharb>how long were atomics live in more than 1-2 browsers?
19:09:23  <gsathya>we know of apps using atomics, so it's not unlikely that this breaks but with some outreach we can fix it
19:09:45  <gsathya>chrome is shipping SABs now
19:10:57  <devsnek>ljharb: that was my thinking lol
19:11:22  <gsathya>(s/can/could, don't want to be too optimistic after the Array.prototype.values circus)
19:12:34  <devsnek>and at least in node, we know no one uses them because we don't have any use for SAB
19:13:37  <Bakkot>related aside: y'all see https://github.com/WICG/web-locks ?
19:15:33  <devsnek>yeah its super cool
19:15:46  <devsnek>i'm working on a pr for node
19:20:08  * Aquaziquit (Quit: Connection closed for inactivity)
19:37:46  * jwaldenjoined
19:50:55  <ljharb>gsathya: apps using atomics that continued to work without atomics being present?
20:12:45  * MylesBorinsquit (Excess Flood)
20:13:02  * MylesBorinsjoined
20:25:39  <devsnek>leobalter: https://github.com/nodejs/node/pull/22844
20:25:49  <devsnek>i actually was thinking the same thing lo
20:41:02  * keithamusquit (Quit: The Lounge - https://thelounge.github.io)
20:41:47  * keithamusjoined
21:46:21  <aklein>ljharb: note that the vast majority of use of atomics is not by directly-authored JS code but by emscripten-generated apps
21:46:55  * keithamusquit (Quit: The Lounge - https://thelounge.github.io)
21:46:56  <aklein>which effectively are like stale versions of frameworks
21:47:02  <aklein>seems like the mootools thing
21:47:37  * keithamusjoined
21:55:52  <devsnek>but still... none of them would have worked anyway
21:56:00  <devsnek>since sharedarraybuffers were disabled
21:56:13  <devsnek>so just patch emscripten codegen and good to go
21:57:03  <aklein>"just patch codegen" doesn't answer the already-generated case
21:57:17  <aklein>(already-generated and orphaned, that is)
21:57:20  <devsnek>already-generated is broken
21:57:29  <aklein>this is what "don't break the web" is about
21:57:36  <aklein>you could say "just update your mootools" or whatever
21:57:46  <devsnek>if we didn't want to break already generated why does "Atomics.wake" not appear in the spec anymore
21:57:48  <aklein>but we try to keep new versions of JS working with the existing web
21:58:09  <aklein>devsnek: well if we find we can't delete it from Chrome and Firefox I suspect it will get back in
21:58:21  <devsnek>annex b life
21:59:27  <aklein>pretty much
22:01:53  <devsnek>is SAB enabled in anything yet
22:02:47  <aklein>it's in Chrome 67+ I think?
22:03:41  <devsnek>like back to being exposed?
22:04:31  <devsnek>thats sad
22:05:01  * Havvyquit (Read error: Connection reset by peer)
22:10:04  <ljharb>aklein: aha, gotcha. that makes sense, thanks.
22:51:03  * basicdaysquit (Quit: I'm out)
22:58:15  * keithamusquit (Quit: The Lounge - https://thelounge.github.io)
22:58:56  * keithamusjoined
23:33:34  * basicdaysjoined