00:25:35  * vivekp_joined
00:35:30  * vivekp_quit (Quit: This computer has gone to sleep)
00:39:09  * vivekp_joined
00:55:22  * bradleymeckquit (Quit: )
01:11:33  * vivekp_quit (Quit: This computer has gone to sleep)
01:41:50  * vivekp_joined
01:46:07  * AtumTquit (Remote host closed the connection)
02:13:48  * vivekp_quit (Quit: This computer has gone to sleep)
02:23:22  * vivekp_joined
02:46:48  * vivekp_quit (Quit: This computer has gone to sleep)
02:50:16  * vivekp_joined
03:58:54  * vivekp_quit (Quit: This computer has gone to sleep)
04:01:34  * vivekp_joined
04:11:00  * vivekp_quit (Quit: This computer has gone to sleep)
04:12:23  * vivekp_joined
04:27:52  * vivekp_quit (Quit: This computer has gone to sleep)
04:30:23  * vivekp_joined
04:41:18  * vivekp_quit (Quit: This computer has gone to sleep)
04:53:48  * vivekp_joined
04:59:46  * vivekp_quit (Quit: This computer has gone to sleep)
05:00:05  * vivekp_joined
05:40:02  * vivekp_quit (Quit: This computer has gone to sleep)
05:44:23  * vivekp_joined
05:50:54  * vivekp_quit (Quit: This computer has gone to sleep)
06:05:14  * vivekp_joined
07:29:43  * xaxxonjoined
07:54:56  * vivekp_quit (Quit: This computer has gone to sleep)
08:03:56  * vivekp_joined
08:58:19  * AtumTjoined
10:42:34  * vivekp_quit (Quit: This computer has gone to sleep)
10:59:11  * plutoniixquit (Quit: Leaving)
11:11:17  * thefourtheyequit (Ping timeout: 276 seconds)
11:11:17  * scottmgquit (Ping timeout: 276 seconds)
11:11:36  * thefourtheyejoined
11:11:40  * scottmgjoined
11:12:08  * akleinquit (Ping timeout: 276 seconds)
11:13:00  * akleinjoined
11:15:57  * xaxxonquit (Ping timeout: 240 seconds)
11:25:08  * mylesborinsquit (Remote host closed the connection)
11:25:44  * mylesborinsjoined
11:29:11  * bradleymeckjoined
11:43:46  <bradleymeck>https://github.com/v8/v8/blob/239dd8124b9afea52f2983d37c7ac63262a9329f/src/profiler/heap-snapshot-generator.cc#L2408-L2415 has a comment that makes it seem unsafe to turn of the forced GC before generating the actual snapshot graph... but I don't want to GC the heap for my random test case... mmm
13:25:00  * vivekp_joined
13:50:28  * bradleymeckquit (Quit: bradleymeck)
13:51:40  * bradleymeckjoined
14:07:19  * bradleymeckquit (Quit: bradleymeck)
14:12:56  * Pablo[m]quit (Ping timeout: 248 seconds)
14:15:19  * Pablo[m]joined
14:46:37  * bradleymeckjoined
16:00:35  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://build.chromium.org/p/client.v8.ports/builders/V8%20Arm/builds/6106 from 954146a5cf923cfa0462f22d4525fbb81b96d761)
16:05:37  <trungl-bot>Tree opened by machenbach@chromium.org: open - flake
16:22:55  * RT|Chatzillaquit (Quit: ChatZilla 0.9.86.1 [Firefox 2.0.0.22pre/2010030309])
16:36:19  * seventhjoined
16:56:11  * gsathyaquit (Quit: WeeChat 1.5)
16:58:09  * gsathyajoined
17:20:14  <trungl-bot>Tree closed by buildbot@chromium.org: closed (https://build.chromium.org/p/client.v8/builders/V8%20Win64%20-%20msvc/builds/1217 from b012816155bb1bb53c3f4fc839b8ba42dbb6af98)
17:43:25  <trungl-bot>Tree opened by machenbach@google.com: open
17:57:00  * bradleymeckquit (Quit: bradleymeck)
19:00:54  * Guest59quit (Quit: My Mac has gone to sleep. ZZZzzz…)
19:03:19  * Guest59joined
19:10:15  * Guest59_joined
19:13:48  * Guest59quit (Ping timeout: 240 seconds)
19:17:58  * bradleymeckjoined
19:25:34  * seventhquit (Remote host closed the connection)
20:10:23  * bradleymeckquit (Quit: bradleymeck)
20:24:27  <cloudshu>can the type feedback vector cause things to leak?
20:27:07  <aklein>cloudshu: I would say...maybe? I don't know where we've swung at the moment WRT weakness of things there
20:27:38  <cloudshu>all the slots look strong to me currently, but i'm like maybe 20% sure on what these BodyDescriptors actually do wrt marking
20:30:22  <aklein>cloudshu: is this still with that rather old version?
20:30:40  <cloudshu>i'm still working off that, but i'm asking about current reality
20:31:50  <aklein>cloudshu: it looks to me like optimized code should only be weakly-held
20:31:57  <aklein>so I'd expect it not to leak
20:32:24  <aklein>cloudshu: not sure about the type feedback itself though
20:33:16  <aklein>cloudshu: my colleague jkummerow tells me that the standard pattern is for type feedback to also be weak
20:33:38  <aklein>though that's not enforced at the FeedbackVector level, but rather by its clients (which should be wrapping things in WeakCells)
20:34:01  <cloudshu>aklein: ah, yeah i'm not familiar with the tracing schema yet
20:34:31  <cloudshu>aklein: i was going off of JSFunction's kFeedbackVectorOffset, which is < kEndOfStrongFieldsOffset, which made me assume it was strong
20:34:39  <cloudshu>aklein: where is the WeakCell wrapping?
20:35:54  <aklein>cloudshu: I'd just search for NewWeakCell
20:36:08  <aklein>lots of these in src/ic/ic.cc for example
20:36:18  <cloudshu>ah ha okay
20:42:57  <cloudshu>ah yeah okay, and in this old version all the FeedbackNexus wraps all the observed stuff in WeakCells
20:43:03  <cloudshu>:+1:
20:47:31  * bradleymeckjoined
20:53:21  * bradleymeckquit (Quit: bradleymeck)
22:09:20  * bradleymeckjoined
22:30:51  * zvquit (Ping timeout: 268 seconds)
22:49:12  * oleavr_joined
22:49:23  * RT|Chatzillajoined
22:50:23  * oleavr_changed nick to oleavr
23:01:46  * Guest59_quit (Quit: My Mac has gone to sleep. ZZZzzz…)
23:02:12  * zvjoined
23:27:20  * Guest59joined
23:29:10  * Guest59quit (Client Quit)
23:30:11  * Guest59joined
23:39:24  * vivekp_quit (Quit: Leaving)
23:45:06  * AtumT_joined
23:48:18  * AtumTquit (Ping timeout: 246 seconds)
23:59:18  * AtumT_quit (Ping timeout: 240 seconds)