00:09:21  * rmgjoined
00:13:45  * rmgquit (Ping timeout: 246 seconds)
01:07:54  * happy-dudequit (Quit: Connection closed for inactivity)
01:52:55  * rgrinbergjoined
03:09:34  * rgrinbergquit (Ping timeout: 252 seconds)
06:22:50  * rmg_quit (Remote host closed the connection)
06:39:14  * pspi_changed nick to pspi
07:23:42  * rmgjoined
07:28:06  * rmgquit (Ping timeout: 250 seconds)
08:09:12  * davijoined
08:24:27  * rmgjoined
08:29:23  * rgrinbergjoined
08:30:13  * rmgquit (Ping timeout: 252 seconds)
08:35:58  * tunniclm_quit (Read error: Connection reset by peer)
08:53:28  * rgrinbergquit (Ping timeout: 244 seconds)
09:26:26  * rmgjoined
09:30:43  * rmgquit (Ping timeout: 252 seconds)
09:39:27  * daviquit (Ping timeout: 244 seconds)
09:48:19  * seishunjoined
10:27:05  * rendarjoined
10:27:31  * rmgjoined
10:32:34  * rmgquit (Ping timeout: 272 seconds)
12:11:40  * davijoined
13:18:48  * seishunquit (Ping timeout: 246 seconds)
13:21:40  * seishunjoined
13:23:09  * daviquit (Ping timeout: 244 seconds)
13:34:17  <seishun>I'm thinking of adding a workaround for https://github.com/nodejs/node/issues/6763 to libuv
13:35:06  <seishun>basically divide the timeout by two if it's greater than some threshold value
13:35:11  <seishun>based on a heuristic that the pc will likely not sleep longer than be awake
13:38:49  <rendar>seishun: hmmm
13:39:36  <rendar>seishun: so basically the timeout we pass to GetQueuedCompletionStatus won't advance while it's sleeping?
13:39:48  <seishun>rendar: exactly
13:40:03  <seishun>but it does advance on windows 7
13:40:32  <rendar>so in windows 7 and < it works fine, it doesn't work in windows >=10
13:40:43  <seishun>I've only tested on windows 7 and 10
13:41:02  <rendar>ok
13:41:19  <rendar>seishun: well, if so, isn't it a Windows bug? we cannot doing anything about it..
13:41:58  <seishun>we can work around it
13:42:06  <rendar>how?
13:42:14  <seishun>the way I described above
13:43:33  <rendar>hmm
13:44:22  <rendar>but how do you set that threshold value? and if you divide it by two, won't the wakeup time be still wrong?
13:46:32  <seishun>if GetQueuedCompletionStatus returns early, libuv just starts the timer for the remaining time
13:48:39  <rendar>seishun: but if you divide the timeout by 2, how do you know that GQCS returned early? i mean, shuuldn't you save the actual time and compare it after, when GQCS returns?
13:49:02  <seishun>rendar: that's what libuv already dos
13:49:03  <seishun>does
13:51:57  <rendar>hmm
13:52:24  <rendar>does libuv do that to solve this bug? or it does that everywhere?
13:53:54  <seishun>it does that to solve another bug, where GQCS sometimes returns early
13:56:00  <rendar>oh..
13:57:08  <rendar>seishun: can you link me the exact code of this precise thing? or at least tell me the file in libuv tree?
13:57:58  <seishun>rendar: https://github.com/libuv/libuv/blob/b12624c13693c4d29ca84b3556eadc9e9c0936a4/src/win/core.c#L288
13:58:53  <rendar>seishun: oh thanks
14:00:06  <rendar>seishun: well, if libuv already implemented this thing of the time to solve another bug, shouldn't this inncidentally solve also the bug you reported, and making libuv your-bug-proof? how did you discover this? by playing with libuv or playing with bare GQCS?
14:01:40  <seishun>rendar: I discovered it by playing with node.js. And no, this workaround doesn't solve the bug where it returns too late
14:02:49  <rendar>oh, because it returns too late, and not too earlier
14:20:30  <saghul>seishun: not sure that would work
14:20:53  <saghul>my Windows 10 machine has been hibernating for a month now :-P
14:22:19  <saghul>FWIW, early returns are the exception, not the norm, I don't think we want to make that the common case
14:22:57  <seishun>saghul: do you have any other suggestions?
14:23:37  <saghul>seishun: nope
14:24:23  <saghul>maybe @ orangemocha, maybe he can ask someone on the Windows team at Microsoft
14:28:02  <seishun>saghul: should I ping him in the node.js issue?
14:28:15  <saghul>seishun: yeah
14:28:25  <saghul>he is in platform-windows, but he might have missed the pint
14:28:28  <saghul>*ping
14:32:05  <seishun>done
14:32:15  <seishun>hope to find out more about this windows 10 weirdness
14:32:29  <seishun>maybe they'll just go ahead and fix it in the next build
16:31:42  * rmgjoined
16:35:28  * seishunquit (Ping timeout: 272 seconds)
16:36:03  * rmgquit (Ping timeout: 240 seconds)
16:51:47  * rgrinbergjoined
16:54:37  * seishunjoined
17:32:37  * rmgjoined
17:36:41  * rmgquit (Ping timeout: 244 seconds)
17:45:18  * davijoined
18:22:37  * Octylpyjoined
18:25:27  <Octylpy>hi. i'm trying to make a simple udp echo server, but it crashes right after calling uv_udp_send. the message sends successfully, but the program crashes right after with EXC_BAD_ACCESS. any help would be gladly appreciated
18:31:26  <nathan7>pastebin your code, for starters
18:32:56  <Octylpy>http://pastebin.com/JXHu6xFp whoops. should have done that in the first place
18:33:22  * rmgjoined
18:37:48  * rmgquit (Ping timeout: 244 seconds)
19:00:13  <saghul>Octylpy: the send request needs to be valid until the callback is called
19:00:21  <saghul>your is allocated on the stack
19:00:29  <saghul>so boom
19:04:56  * seishunquit (Ping timeout: 272 seconds)
19:05:41  * happy-dudejoined
19:13:42  <Octylpy>saghul: ohhh ok. that fixed it. thanks
19:16:59  * seishunjoined
19:26:52  * Octylpyquit (Quit: Page closed)
19:28:27  * daviquit (Ping timeout: 258 seconds)
19:34:06  * rmgjoined
19:38:33  * rmgquit (Ping timeout: 246 seconds)
20:34:54  * rmgjoined
20:37:45  * rgrinbergquit (Ping timeout: 276 seconds)
20:39:22  * rmgquit (Ping timeout: 250 seconds)
21:12:14  * seishunquit (Ping timeout: 272 seconds)
21:17:54  * happy-dudequit (Quit: Connection closed for inactivity)
21:31:13  * rendarquit (Ping timeout: 252 seconds)
21:35:32  * rmgjoined
21:40:06  * rmgquit (Ping timeout: 272 seconds)
21:49:10  * tunniclm_joined
22:00:45  * rendarjoined
22:04:49  * srl295quit (Quit: No Ping reply in 180 seconds.)
22:05:55  * srl295joined
22:05:56  * srl295quit (Changing host)
22:05:56  * srl295joined
22:50:27  * thefourtheye___quit (Quit: Connection closed for inactivity)
23:30:44  * tunniclm_quit (Read error: Connection reset by peer)
23:50:13  * rendarquit (Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!)