00:03:24  <sandro>this is so unlikely though
00:03:41  <indutny>(shrug)
00:09:30  <sandro>indutny: this is not the case
00:09:47  <sandro>the handle is embedded in an object that is still alive, i checked the address
00:09:57  <indutny>well, valgrind should certainly print something before it crashes
00:10:01  <indutny>did it?
00:10:10  <indutny>asan would probably do it too
00:10:21  <sandro>yes, but it only shows the very same reason why it crashes (dereferencing 0x8)
00:10:25  <sandro>yes
00:10:27  <sandro>same thing
00:10:47  <sandro>if i let it core normally, if i let it crash under asan, if i let it get caught by valgrind
00:10:57  <sandro>it's always at the very same moment
00:11:07  <sandro>asan and valgrind don't see any issue up until then
00:11:47  <sandro>indutny: can that "heap" data structure have been corrupted?
00:13:17  <indutny>it appears that it is
00:13:30  <indutny>please file a bug on github with the data that you have
00:13:33  <indutny>we'll take a look
00:13:39  <indutny>btw, which libuv version are you using?
00:15:34  <sandro>1.19.2
00:15:48  <sandro>indutny: i am not saying that it's getting corrupt because of your fault
00:16:02  <sandro>it could be my code
00:16:11  <sandro>(and most likely it is)
00:16:14  <indutny>there's one thing that you could try
00:16:15  <indutny>if you have time
00:16:21  <indutny>read/write breakpoints in gdb
00:16:23  <sandro>indutny: sure, i have time
00:16:32  <indutny>just to see what code writes there
00:16:40  <sandro>indutny: is there a way to introduce them programmatically?
00:16:42  <indutny>this assumes that you know the structure ahead of time
00:16:46  <indutny>ah, so you don't
00:16:48  <indutny>okay then
00:16:55  <sandro>i.e.: the compiler emits them and then i run under gdb and they trigger?
00:17:05  <indutny>I don't think there is a way to do it programmatically
00:17:18  <indutny>maybe I'm wrong, though
00:20:00  <sandro>i could set a regular breakpoint at object creation and then my watchpoint
00:20:02  <sandro>https://web.archive.org/web/20130605011618/http://www.technochakra.com:80/software-breakpoints/
00:25:01  <indutny>possibly
00:27:37  <sandro>i need to go run some errands soon though
00:27:49  <sandro>but i'll keep you posted on what i discover
00:29:46  <indutny>thanks!
00:46:12  * listenmorequit (Remote host closed the connection)
00:46:41  * listenmorejoined
01:19:45  * AtumTquit (Remote host closed the connection)
05:16:41  * toxedviruschanged nick to txdv
07:14:15  * sandroquit (Ping timeout: 260 seconds)
09:57:21  * niskaquit (Quit: Leaving)
10:00:15  * erikj-quit (Quit: kfonx)
10:00:39  * erikj_joined
10:02:10  * listenmorequit (Remote host closed the connection)
10:02:43  * listenmorejoined
10:03:24  * niskajoined
10:25:12  * mylesborinsquit (Quit: farewell for now)
10:25:41  * mylesborinsjoined
12:16:29  * chhmmljoined
12:16:30  * chhmmlquit (Client Quit)
15:48:12  * lukaquit (Ping timeout: 260 seconds)
15:55:26  * bengl_joined
16:03:03  * benglquit (*.net *.split)
16:03:03  * bzozquit (*.net *.split)
16:03:07  * bengl_changed nick to bengl
16:10:51  * bzozjoined
17:57:48  * listenmorequit (Quit: Konversation terminated!)
17:58:05  * listenmorejoined
18:29:34  * joocain2quit (Remote host closed the connection)
18:29:54  * joocain2joined
18:44:43  * AtumTjoined
19:35:47  * joocain2quit (Remote host closed the connection)
19:36:11  * joocain2joined
22:02:01  * listenmorequit (Remote host closed the connection)
22:02:37  * listenmorejoined
22:07:23  * listenmorequit (*.net *.split)
22:07:23  * AtumTquit (*.net *.split)
22:07:25  * keegansquit (*.net *.split)
22:07:26  * duvnellquit (*.net *.split)
22:07:26  * robertkowalskiquit (*.net *.split)
22:07:53  * listenmorejoined
22:07:53  * AtumTjoined
22:07:53  * duvnelljoined
22:07:53  * keegansjoined
22:07:53  * robertkowalskijoined
22:10:25  * listenmorequit (Read error: Connection reset by peer)
22:11:00  * listenmorejoined
22:19:44  * listenmorequit (Remote host closed the connection)
22:20:13  * listenmorejoined
23:16:14  * sandro_joined
23:30:02  <sandro_>hi
23:31:22  <sandro_>indutny: I was wondering... I have most of my handles as "intrusive" members of my connection objects. Couldn't I simply assert() in the object destructor that the handles have all been properly closed?
23:56:16  <txdv>node context?