|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:30
lucasb left
00:38
pamplemousse joined
00:48
pamplemousse left
02:32
zakharyas joined
02:39
zakharyas left
03:03
zakharyas joined
03:58
pamplemousse joined
05:11
brrt joined
05:26
pamplemousse left
05:57
robertle left
07:03
domidumont joined
07:20
robertle joined
07:35
brrt left
07:50
Kaiepi left
07:53
Kaiepi joined
09:29
zakharyas left
10:15
brrt joined
10:25
ilogger2 joined
10:26
eater joined,
nebuchadnezzar joined,
tadzik joined,
nebuchadnezzar is now known as Guest90858
10:29
bartolin joined,
avar joined,
avar left,
avar joined,
japhb joined
10:33
Voldenet joined,
Voldenet left,
Voldenet joined
11:54
brrt joined
|
|||
| brrt | \o | 12:25 | |
| nwc10 | o/ | ||
| moritz | |o| | 12:26 | |
| timotimo | …o… | 12:42 | |
| tadzik | ᕕ( ᐛ )ᕗ | 12:43 | |
| jnthn | ~~\o/~🦈~ | 12:50 | |
| timotimo | nooo jnthn don't drown | 12:52 | |
| though we have recently clarified that drowning people don't wave their arms | 12:53 | ||
| so i guess jnthn's just having safe fun in the pool | |||
| tadzik | well, if you're a scuba diver then waving your hands means "HELP" | 12:54 | |
| but maybe jnthn is just enthusiastic to see the shark :) | |||
| timotimo | or, if you're still on land, it means "jazz hands!" | ||
| moritz | safe fun in the pool with sharks? | ||
| sounds legit, kinda | 12:55 | ||
| brrt | I seem to recall that 'most' sharks weren't all that dangerous | 13:02 | |
| esp. not compared to killer whales | |||
| moritz | that's correct | ||
| timotimo | gfycat.com/abandonedtemptingchuckwalla-dog | 13:03 | |
| moritz | en.wikipedia.org/wiki/Shark#Relati...ith_humans "average number of fatalities worldwide per year between 2001 and 2006 from unprovoked shark attacks is 4.3" | ||
| "Contrary to popular belief, only a few sharks are dangerous to humans. Out of more than 470 species, only four have been involved in a significant number of fatal, unprovoked attacks on humans" | |||
| jnthn | m: say 4/470 | 13:05 | |
| evalable6 | 0.008511 | ||
| jnthn | Seems my chances or survival are pretty decent :P | ||
|
13:19
evalable6 joined
|
|||
| timotimo | gist.github.com/timo/a2f4bb66f0811...9f14ef75f0 | 13:45 | |
| i might also want to put info about "how many GCs have been major so far" in the heap snapshots | 13:46 | ||
| snapshot number and gc seq num don't relate if you've got a confprog set up | |||
|
13:51
Guest90858 is now known as nebuchadnezzar
13:56
lucasb joined
|
|||
| Guest71418 | .seen nine | 13:57 | |
| tellable6 | Guest71418, I saw nine 2019-08-30T12:05:51Z in #perl6-dev: <nine> That's...a little bit surprising | ||
| Kaiepi | timotimo, you have time to look at my code today? | 14:51 | |
| timotimo | i'll try! | ||
| i just had enough time to make a change to JSON::Fast | 14:52 | ||
| Fix async UDP sockets, pass socket types and protocols when resolving - this is the commit i should be at? | 14:58 | ||
| Kaiepi: can you give me the test code again? can't find it in my irc history right now | 15:00 | ||
| Kaiepi | perl6 -e 'my IO::Socket::Async $s .= bind-udp: "127.0.0.1", 8000; my Tap $t = $s.Supply.tap({ .print }); my IO::Socket::Async $c .= udp; await $c.print-to: "127.0.0.1", 8000, "ayy lmao"' | 15:02 | |
| timotimo | |||
| yeah, that commit and the one before it | |||
| first one doesn't change anything related to async udp sockets | 15:03 | ||
| timotimo | OK | ||
| "invalid argument" is the error i should be seeing? | 15:04 | ||
| Kaiepi | yep | 15:14 | |
| the other kind is it claiming the socket is closed when it isn't, that's harder to debug with since you can't break on uv_send_udp though | 15:15 | ||
| though i guess you could just break on uv_is_closing instead, that would be the only place it gets used at all in the program | 15:16 | ||
| timotimo | rr is refusing my hardware watchpoints wtf | 15:17 | |
| Kaiepi | how do you use rr? there's no command for it in gdb and if it's a program, it doesn't exist on my machine | 15:18 | |
| timotimo | "rr record perl6 blah blah", then "rr replay" | ||
| it's an external program indeed | |||
| rr-project.org or so | |||
|
15:20
patrickb joined
|
|||
| timotimo | it looks less like the handle data gets overwritten with zeroes | 15:25 | |
| and more like it's never written | |||
| like, i follow the memory inside that and it literally turns from "this memory isn't mapped" to the value it has when the code that errors runs | 15:27 | ||
|
15:27
pamplemousse joined
|
|||
| timotimo | also | 15:28 | |
| the code runs as response to a IO::Socket::Async::SocketCancellation being worked on | |||
| that seems wrong? | |||
| OK, the type there isn't as important | 15:32 | ||
| in write_bytes_to_async you may want MVM_is_null instead of == NULL | 15:35 | ||
| that also catches VMNull repr'd objects | |||
| ah, ok, write_bytes_to checks for concreteness | 15:37 | ||
|
15:51
brrt left
|
|||
| timotimo | Kaiepi: are you here? | 15:59 | |
| Kaiepi: my latest finding is that do_setup_setup is called with an ssi that has no bind_address at all, so uv_udp_init is never called, that's potentially why the whole handle is nulled out | 16:08 | ||
| Kaiepi | ahhh | 16:12 | |
| that'd explain it | |||
| timotimo | and since ssi->error was set to 0 in setup_setup, the error path isn't taken | ||
| Kaiepi | since before it was only ever using one address, uv_udp_init was still getting called before because it only cared if ssi->bind_addr was NULL when binding the socket. with the loop there, that's no longer the case | 16:17 | |
| timotimo | right | 16:18 | |
| oh, also, you may want to malloc your char *hostname with more than just sizeof(char) | 16:19 | ||
| because that'll give you a 1 byte big buffer, rounded up to whatever malloc's minimum allocation size is | |||
| Kaiepi | MVM_calloc(MAX_HOST, sizeof(char)) or whatever? | 16:20 | |
| timotimo | *shrug* | ||
| it's only debug code, but it should still maybe not immediately buffer overflow :D | |||
| Kaiepi | i'm guessing it only ever worked because inet_ntop reallocs it or something | 16:22 | |
| timotimo | aha, the last argument to inet_ntop is the size of your dst buffer | 16:25 | |
| Kaiepi | i'm passing the wrong thing then, i thought it was the size of the struct sockaddr you're passing it | 16:26 | |
|
16:48
pamplemousse left
|
|||
| Kaiepi | awesome, everything works now! thanks timotimo | 17:12 | |
| timotimo | i accept tips, LOL | ||
| BBL | |||
|
17:35
harrow left
17:36
harrow joined
18:06
robertle joined
18:10
Kaiepi joined
18:13
Ven`` joined
18:14
Ven`` left
19:37
Ven`` joined
19:52
Ven`` left
|
|||
| patrickb | I'm currently looking at github.com/MoarVM/MoarVM/blob/mast...ops.c#L255 | 20:51 | |
| Some quick tests seem to indicate this ignores the root user. | |||
| Is there something obvious I miss or should this code also test for euid() == 0? | 20:55 | ||
| I got there by tracking down github.com/rakudo/rakudo/issues/3090 In that bugreport people are suggesting libuv broke between 1.24.1 and 1.30.1. | 20:57 | ||
|
21:04
Ven`` joined
21:10
Ven`` left
21:11
Ven`` joined,
Ven`` left
|
|||
| patrickb | But from my tests and looking at the code I'd suspect there is a check for root missing. Opinions? | 21:22 | |
|
22:32
patrickb left
22:40
evalable6 left
22:43
evalable6 joined
22:46
lucasb left
22:47
vesper11 joined
23:16
Kaiepi left
|
|||