[00:23] *** rmv left [01:53] *** kylese left [01:53] *** kylese joined [01:54] *** simcop2387 left [01:54] *** simcop2387 joined [01:56] *** MasterDuke joined [02:15] *** kylese left [02:15] *** kylese joined [02:37] *** guifa_ left [02:45] *** MasterDuke left [02:50] *** coverable6 left [02:50] *** unicodable6__ left [02:50] *** sourceable6 left [02:50] *** quotable6 left [02:50] *** shareable6 left [02:50] *** greppable6 left [02:50] *** tellable6 left [02:50] *** evalable6 left [02:50] *** committable6 left [02:50] *** linkable6 left [02:50] *** nativecallable6 left [02:50] *** huggable6 left [02:50] *** notable6 left [02:50] *** releasable6 left [02:50] *** bisectable6 left [02:50] *** bloatable6 left [02:50] *** benchable6 left [02:53] *** committable6 joined [02:53] *** coverable6 joined [02:53] *** greppable6 joined [02:54] *** unicodable6 joined [02:54] *** bloatable6 joined [02:54] *** nativecallable6 joined [02:54] *** benchable6 joined [02:54] *** bisectable6 joined [02:55] *** notable6 joined [02:55] *** quotable6 joined [02:55] *** linkable6 joined [02:55] *** releasable6 joined [02:55] *** shareable6 joined [02:56] *** evalable6 joined [02:56] *** sourceable6 joined [02:56] *** huggable6 joined [02:56] *** tellable6 joined [03:05] Voldenet: I did try the interval supply, which stops, but not sure if that’s cpu speed issue because the other supply even in the same thread continues to work.. but indeed I didn’t take into account ipv6 but worst case it should “simply” make the connection not work since the ip:port is wrong but not block the rest of the program to work [03:07] When the freeze happen, the tracker thread continues but get stuck at the http request step, the main thread continues to disconnect clients with timeout but that’s about it, even tho I added interval in the main file to show progress every 5s, it doesn’t do it, but that’s in the same thread so if one work, the others should work too.. ? [03:23] *** vasko4535586 left [03:42] Cro::HTTP::Client has no timeout by default [03:44] or rather, it has timeout on connection and headers, but not body [03:44] by default [03:44] https://cro.raku.org/docs/reference/cro-http-client#Timeouts [03:45] so perhaps tracker gets connected but fails for some reason [03:54] > use Cro::HTTP::Client; react whenever Cro::HTTP::Client.get($!uri, :timeout(10)) { whenever .body-blob { .say } } # consider this [04:11] *** tejr left [04:17] *** belluzj joined [04:34] *** [dpk] joined [04:34] *** inspork_ joined [04:35] *** releasable6 left [04:35] *** linkable6 left [04:35] *** coverable6 left [04:35] *** unicodable6 left [04:35] *** committable6 left [04:35] *** tellable6 left [04:35] *** evalable6 left [04:35] *** nativecallable6 left [04:35] *** benchable6 left [04:35] *** quotable6 left [04:35] *** notable6 left [04:35] *** dpk left [04:35] *** inspork left [04:35] *** coverable6 joined [04:35] *** releasable6 joined [04:35] *** linkable6 joined [04:35] *** inspork_ is now known as inspork [04:40] simon_sibl: Are you using Cro::HTTP::Client? If so, can you create an unencrypted TCP dump of the lock up? [04:40] *** tejr joined [04:44] `start` is extremely good at covering up errors, so maybe it just dies [04:44] used to be, at least [04:47] what I'd do is something similar to this https://pastebin.com/GpBAG0vk [04:48] see https://github.com/croservices/cro-http/pull/162 [04:48] for how to do that [05:01] i.e. a normal TCP dump dump + the ssl session key log [05:02] That then allows looking at the http procol. [05:05] it's this piece of code, I think it connects to trackers from the file in that repo https://github.com/4zv4l/rbt/blob/main/lib/BT/Tracker.rakumod [05:09] still, wouldn't timeout catch tls hangs? [05:27] *** Sgeo left [05:27] *** Sgeo joined [06:18] *** rmv joined [06:18] *** rmv left [06:18] *** rmv joined [06:23] *** rmv left [07:18] *** rmv joined [07:18] *** rmv left [07:18] *** rmv joined [07:24] *** eseyman left [07:31] *** manu_ joined [07:32] *** manu_ is now known as eseyman [07:36] *** [dpk] is now known as dpk [07:59] *** kylese left [08:00] *** kylese joined [10:12] Not necessarily. If there is a bug in Cro's HTTP (2) stack (and there have been quite a few in the past) then a timeout most probably doesn't help. [10:42] *** guifa_ joined [10:50] *** belluzj left [10:51] *** Sgeo left [12:49] *** guifa_ left [12:59] *** guifa_ joined [13:30] only one typo in my talk this year [13:30] lol [13:50] *** belluzj joined [13:52] *** belluzj left [13:55] *** guifa_ left [14:35] *** guifa_ joined [14:46] *** shareable6 left [14:46] *** huggable6 left [14:46] *** sourceable6 left [14:46] *** bloatable6 left [14:46] *** linkable6 left [14:46] *** bisectable6 left [14:46] *** coverable6 left [14:46] *** releasable6 left [14:46] *** greppable6 left [14:48] *** shareable6 joined [14:49] *** greppable6 joined [14:49] *** bloatable6 joined [14:49] *** nativecallable6 joined [14:49] *** quotable6 joined [14:49] *** linkable6 joined [14:50] *** unicodable6 joined [14:50] *** sourceable6 joined [14:50] *** huggable6 joined [14:50] *** tellable6 joined [14:50] *** releasable6 joined [14:51] *** bisectable6 joined [14:51] *** benchable6 joined [14:51] *** notable6 joined [14:51] *** committable6 joined [14:51] *** coverable6 joined [14:52] *** evalable6 joined [15:12] *** abraxxa joined [15:14] *** abraxxa left [15:15] *** abraxxa joined [15:28] *** guifa_ left [16:26] *** guifa_ joined [16:47] m: role N[::TValue] { has TValue @!x; }; N[Int].new; say 42 [16:47] rakudo-moar 5c15283e5: OUTPUT: «42␤» [16:47] Voldenet ^ (from https://github.com/rakudo/rakudo/issues/6056) [17:56] *** wayland76 joined [17:56] *** wayland left [18:20] *** guifa_ left [18:29] *** guifa_ joined [18:32] *** MasterDuke joined [18:34] *** guifa_ left [19:15] *** Sgeo joined [19:34] *** MasterDuke left [20:27] m: role N[::TValue] { has TValue @!x[$n]; }; N[Int].new [20:27] rakudo-moar 5c15283e5: OUTPUT: «===SORRY!=== Error while compiling ␤Variable '$n' is not declared. Perhaps you forgot a 'sub' if this was␤intended to be part of a signature?␤at :1␤------> role N[::TValue] { has TValue @!x[$n]; }; N[Int].new␤» [20:27] m: my $n = 42; role N[::TValue] { has TValue @!x[$n]; }; N[Int].new [20:27] rakudo-moar 5c15283e5: ( no output ) [20:28] yeah, it sounds like it's all fixed [20:37] *** vrurg joined [20:39] *** rmv left [20:39] *** vrurg_ left [20:40] *** rmv joined [20:40] *** rmv left [20:40] *** rmv joined [21:07] *** MasterDuke joined [21:29] *** MasterDuke left [22:08] *** abraxxa left [22:39] *** AlexDaniel joined [22:42] *** bisectable6 left [22:42] *** bloatable6 left [22:42] *** unicodable6 left [22:42] *** notable6 left [22:42] *** tellable6 left [22:42] *** releasable6 left [22:42] *** huggable6 left [22:42] *** linkable6 left [22:42] *** sourceable6 left [22:42] *** quotable6 left [22:42] *** evalable6 left [22:42] *** benchable6 left [22:42] *** shareable6 left [22:42] *** coverable6 left [22:42] *** nativecallable6 left [22:42] *** committable6 left [22:46] *** greppable6 left [22:48] *** unicodable6 joined [22:48] *** linkable6 joined [22:49] *** evalable6 joined [22:49] *** shareable6 joined [22:49] *** bloatable6 joined [22:49] *** greppable6 joined [22:50] *** coverable6 joined [22:50] *** tellable6 joined [22:50] *** sourceable6 joined [22:50] *** benchable6 joined [22:51] *** nativecallable6 joined [22:51] *** releasable6 joined [22:51] *** committable6 joined [22:51] *** quotable6 joined [22:51] *** bisectable6 joined [22:51] *** notable6 joined [22:51] *** huggable6 joined [23:58] patrickb: I misunderstood your problem. When I carefully read timo's followup, I understood it better, and think that the solutions that others have offered are the better ones for your situation.