00:39:52  * bnoordhuisquit (Ping timeout: 245 seconds)
01:19:27  * plutoniixquit (Ping timeout: 245 seconds)
01:38:14  * jmar777joined
01:45:51  * plutoniixjoined
01:46:17  * bnoordhuisjoined
01:50:49  * bnoordhuisquit (Ping timeout: 255 seconds)
02:10:13  * plutoniixquit (Quit: จรลี จรลา)
02:16:43  * petka__quit (Quit: Connection closed for inactivity)
04:06:37  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "compile" on http://build.chromium.org/p/client.v8/builders/V8-Blink%20Mac/builds/2076 "V8-Blink Mac" from fb55eae323e69e6382efb7afda3301776a4206fd: [email protected])
04:10:31  * caitp-joined
04:43:08  <trungl-bot>Tree opened by [email protected]: Tree is open (someone with debug fuu should fix the ProcessDebugMessages flakiness)
05:47:51  * caitp-quit (Ping timeout: 256 seconds)
06:01:29  <mathiasbynens>getting this on OS X – any ideas?
06:01:32  <mathiasbynens>$ make native
06:01:32  <mathiasbynens>make: *** No rule to make target `buildtools/third_party/libc++abi/libc++abi.gyp', needed by `out/Makefile.native'. Stop.
06:02:43  * mathiasbynensruns `gclient sync`
06:45:50  * jmar777quit (Remote host closed the connection)
07:14:06  * mostynbjoined
07:18:09  * enaqxquit (Ping timeout: 245 seconds)
07:22:29  * juanjosanchezquit (Ping timeout: 256 seconds)
07:23:56  * muellijoined
07:31:29  * mostynbquit (Remote host closed the connection)
07:32:24  * mostynbjoined
09:05:50  * Lethalmanjoined
09:56:37  * juanjosanchezjoined
10:04:29  * juanjosanchez_joined
10:06:57  * juanjosanchezquit (Ping timeout: 245 seconds)
10:11:00  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "compile" on http://build.chromium.org/p/client.v8/builders/V8%20Arm%20-%20builder/builds/2632 "V8 Arm - builder" from bd61a85fafd6461a40dc1e20252fd843f148e837: [email protected])
10:11:59  * juanjosanchez_quit (Remote host closed the connection)
10:15:22  * ashnurquit (Remote host closed the connection)
10:17:42  * ashnurjoined
10:28:00  * juanjosanchez_joined
10:34:16  * juanjosanchez_quit (Quit: Ex-Chat)
10:34:42  * petka__joined
10:36:11  * bnoordhuisjoined
10:36:42  * juanjosanchezjoined
10:47:40  * muelliquit (Ping timeout: 255 seconds)
10:52:43  * juanjosanchez_joined
10:55:33  * juanjosanchezquit (Ping timeout: 256 seconds)
11:26:08  <trungl-bot>Tree closed by [email protected]: Tree is closed (fixing...)
11:27:25  * dobsonquit (Read error: Connection reset by peer)
11:34:31  * dobsonjoined
11:37:20  <trungl-bot>Tree closed by [email protected]: Tree is closed (potential fix in flight)
11:47:29  * juanjosanchez_quit (Read error: No route to host)
11:47:50  * juanjosanchez_joined
11:54:18  * juanjosanchez_quit (Read error: Connection reset by peer)
11:54:34  * juanjosanchez_joined
11:55:38  <trungl-bot>Tree opened by [email protected]: Tree is open (potential fix in flight)
12:03:12  * juanjosanchez_quit (Ping timeout: 245 seconds)
12:16:46  * plutoniixjoined
12:24:44  * KillerJimjoined
13:46:19  * caitp-joined
13:48:22  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "Check" on http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm64%20-%20sim%20-%20debug/builds/1754 "V8 Linux - arm64 - sim - debug" from 9159c419e8773d939c899989af3851edd33312a3: [email protected],[email protected],[email protected],[email protected])
14:00:42  * enaqxjoined
14:28:53  * caitp-quit (Ping timeout: 246 seconds)
14:39:08  <trungl-bot>Tree opened by [email protected]: Tree is open
14:42:25  * caitp-joined
14:43:11  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "Check" on http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20isolates/builds/2234 "V8 Linux - isolates" from cb6ea146dc8813f7dc8fa7445d6d0f542509f317: [email protected])
14:44:05  * caitpquit (Ping timeout: 244 seconds)
14:48:17  <trungl-bot>Tree opened by [email protected]: Tree is open (it was a revert, so the test looks flaky)
14:59:33  * juanjosanchez_joined
15:04:22  * juanjosanchez_quit (Ping timeout: 265 seconds)
15:06:44  * jmar777joined
15:26:56  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "Check" on http://build.chromium.org/p/client.v8/builders/V8%20Linux64%20TSAN/builds/2523 "V8 Linux64 TSAN" from 30c71859edceae15d299534c0aecfb0362b94434: [email protected],[email protected])
15:30:03  * caitp__joined
15:33:31  <bnoordhuis>out of curiosity, what's the plan / goal for 'use strong'?
15:34:32  <caitp__>my understanding is, opportunities to eliminate some things that are hard to optimize
15:34:39  <caitp__>and that's basically it
15:35:52  <bnoordhuis>is it something that came out of tc39 or is it a v8 special?
15:36:43  * KillerJimquit (Quit: Leaving)
15:37:14  <caitp__>it's all v8, although spidermonkey has done some similar things
15:37:19  <caitp__>like disallowing arguments when rest parameters are used
15:38:09  <caitp__>wonder if that's fixed yet
15:39:11  <bnoordhuis>that reminds me, have you seen this? https://github.com/erights/quasiParserGenerator
15:39:31  * juanjosanchezjoined
15:39:58  <caitp__>I saw the es-discuss message about it saying "please deprecate this code", haven't really looked at it
15:41:16  <caitp__>interesting, but i would predict that performance would be really bad
15:41:35  <bnoordhuis>yes? how so?
15:41:57  <caitp__>tagged templates themselves are rather expensive in v8, although probably a lot faster in SM
15:42:20  <caitp__>that could change in the future, but the caching mechanism is really non-ideal right now
15:42:37  <bnoordhuis>oh, implementation-wise you mean. that just requires the SMOP
15:44:37  <caitp__>right now, the parser builds an array literal featuring the strings, another array featuring the raw strings, and adds each evaluated expression value as an argument --- then it does a lookup for the callsite based on a hash key and comparison of raw string literals with the cached literals
15:44:57  <caitp__>so if you had a lot of different cached template callsites, it would get nasty
15:45:02  <caitp__>and it's pretty bad even without
15:46:33  <bnoordhuis>right. i remember that discussion from one of your CLs
15:47:01  * mostynbquit (Quit: Leaving)
15:47:38  <caitp__>i'd like to see how fast FF's tagged templates are, it's probably a lot quicker and it would take some work to be that fast
15:47:54  <bnoordhuis>what's the most expensive part? evaluating the expressions or doing the lookup?
15:48:29  <bnoordhuis>OSE-less expressions could be evaluated as needed. perhaps easier said than done :)
15:48:59  <caitp__>building the arrays, calculating the hash, Map lookup, iterating over collided hashes to find an exact match even if there is only one cached entry with the hash
15:49:23  <caitp__>and then, if cache lookup fails, adding a new entry
15:49:34  <caitp__>not as simple as you'd like it to be
15:49:59  <bnoordhuis>there was a reason for not embedding it in the code, wasn't there? gc?
15:50:31  <caitp__>the complicated thing is that, templates with the same set of identical raw strings have to use the same callsite object
15:50:43  <caitp__>if you could just build a new one every time it would be easier to optimize
15:51:05  <caitp__>and likely better on the gc, since they wouldn't be tenured
15:51:14  <bnoordhuis>right, that sounds familiar. that's from the spec, isn't it?
15:51:18  <caitp__>yeah
15:51:32  <bnoordhuis>do you know what the rationale for that was?
15:51:58  <caitp__>I haven't read the notes on that particular decision
15:52:16  <bnoordhuis>you just implemented it :)
15:52:34  <caitp__>https://github.com/rwaldron/tc39-notes/blob/master/es6/2014-11/nov-18.md#48-template-literal-call-site-object-caching covered here
15:52:47  <caitp__>yeah, I think arv wanted to get it clarified because the spec was sort of vague
15:53:00  <caitp__>and it wound up being "cache per-realm, per-raw strings"
15:55:07  <bnoordhuis>hrm, i don't think i understand the performance arguments that are being made in those notes. why are 5 call sites worse than 1?
15:56:52  <caitp__>I think it depends on a lot of things --- if you end up with tenured callsites, that contributes to the "big, expensive GC" that hurts performance
15:57:00  <caitp__>so 5 tenured callsites is worse than 1 tenured callsite
15:57:11  <caitp__>but if you're careful with them and they aren't tenured, you do a lot better
15:57:48  <caitp__>but I guess it also enables people to like, special case specific callsites
15:57:50  <caitp__>or similar
15:58:54  <caitp__>i'm not personally convinced by the whole caching thing, it's likely a lot cheaper without
15:59:03  <bnoordhuis>right, i see the gc argument. but why does tc39 mandate that when it can be left to implementors to decide whether they want to do sharing or not?
15:59:16  <bnoordhuis>intuitively, it feels like an implementation detail, not a spec requirement
15:59:19  <caitp__>it's observable in some ways, so
15:59:31  <caitp__>ideally, it wouldn't be observable
16:00:14  <caitp__>since the callsite objects are frozen, you can't really do anything wiht them, so overriding their equality checks based on a hidden hash key seems fine
16:00:24  <bnoordhuis>so, something for es7? :)
16:00:26  <caitp__>until collisions come in
16:01:15  * caitp__changed nick to caitp
16:03:19  <caitp>anyways, I guess they're convinced that cache lookup can be fast, and it probably can
16:03:26  <caitp>just not the way it's implemented in v8
16:12:51  * RT|Chatzillaquit (Quit: ChatZilla 0.9.86.1 [Firefox 2.0.0.22pre/2009081014])
16:25:52  <trungl-bot>Tree opened by [email protected]: Tree is open
16:28:28  * gf3quit (Ping timeout: 265 seconds)
16:29:55  * chrisdickinsonquit (Ping timeout: 265 seconds)
16:31:00  * chrisdickinsonjoined
16:31:12  * gf3joined
16:42:50  * caitpquit (Ping timeout: 252 seconds)
16:49:51  * caitpjoined
16:53:52  * juanjosanchezquit (Ping timeout: 240 seconds)
17:09:08  * trungl-botquit (Ping timeout: 245 seconds)
17:11:20  * trungl-botjoined
17:14:08  <trungl-bot>Tree closed by [email protected]: Tree is closed (Automatic: "compile" on http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20shared/builds/2408 "V8 Linux - shared" from 515d26063468c392db6ad8efc5cf267a6556c8a5: [email protected],[email protected])
17:15:10  <trungl-bot>Tree closed by [email protected]: Tree is closed (ishell is fixing)
17:40:33  <trungl-bot>Tree opened by [email protected]: Tree is open
17:56:26  * Lethalmanquit (Quit: Sto andando via)
18:43:16  * juanjosanchezjoined
18:58:39  * juanjosanchezquit (Remote host closed the connection)
19:15:55  * juanjosanchezjoined
19:42:12  * juanjosanchezquit (Ping timeout: 244 seconds)
19:44:02  * KillerJimjoined
20:07:48  * petka__quit (Ping timeout: 252 seconds)
20:14:22  * petka__joined
20:23:47  * ashnurquit (Ping timeout: 246 seconds)
21:17:05  * juanjosanchezjoined
21:20:01  * juanjosanchezquit (Client Quit)
21:20:49  * russfrankchanged nick to rf
21:20:55  * rfquit (Changing host)
21:20:55  * rfjoined
21:32:25  * jmar777quit (Remote host closed the connection)
21:32:58  * caitpquit (Ping timeout: 265 seconds)
22:29:23  * RT|Chatzillajoined
22:34:55  * KillerJimquit (Quit: gone cubing.. with the wife)
23:17:03  * jmar777joined
23:42:35  * bnoordhuisquit (Ping timeout: 246 seconds)