|
00:07
stanrifkin left
00:19
manu_ joined,
manu_ is now known as eseyman
00:21
itaipu joined
01:28
sibl joined
01:30
librasteve_ left
|
|||
| grondilu | any idea why `sub MAIN { start do-stuff-for-less-than-ten-seconds(); sleep 10; }` would not behave or even segfault? | 02:40 | |
| when `sub MAIN { await start do-stuff-for-less-than-ten-seconds(); }` seems to work just fine? | |||
|
02:42
hulk joined
02:43
kylese left
|
|||
| Voldenet | because it's ran on thread scheduler | 02:45 | |
| and when doing start you can't really expect any kind of timing | |||
| `Thread.start: &do-stuff-for-less-than-ten-seconds;` would start the actual thread in the background that would "hang" the app until aborted | 02:47 | ||
|
03:05
arkiuat left
03:15
hulk left,
kylese joined
|
|||
| grondilu | IIUC what I really want to do is `sub MAIN { my $job = start do-this(); do-that(); await $job; }` | 03:16 | |
| grondilu checks | |||
| still not quite working as expected. | 03:17 | ||
|
03:19
arkiuat joined
|
|||
| grondilu | What I'm trying to do is something like: `sub MAIN { my $graphics-loop = start graphics-loop(); sleep 10; await $graphics-loop; }` | 03:21 | |
| `sleep` here is just meant to simulate some other job. | |||
| gist.github.com/grondilu/8628e20fb...c8aa908a19 | |||
| this doesn't crash anymore, but the graphics loop does not seem to start until the sleep command has terminated. | 03:22 | ||
|
03:27
vasko453558 left
03:33
arkiuat left
03:34
vasko453558 joined
03:49
arkiuat joined
03:56
arkiuat left
04:17
arkiuat joined
04:26
lichtkind_ joined
04:29
lichtkind left
05:10
oodani left
05:13
oodani joined
05:19
ShimmerFairy left
05:30
ShimmerFairy joined
|
|||
| Voldenet | `my $thread = Thread.start(…); $thread.join` would use OS-level scheduler which is probably more predictable in this case, task scheduler is better for tasks that wait more than compute (networking, or interacting with other processes) also `init-white` will malloc the color every time you call it, you'll want to cache it | 06:06 | |
| rarely starting a thread is a good idea, but for any realtime rendering it's actually good, because you can increase priority | 06:07 | ||
| or bind specific thread to the specific cpu core | |||
|
06:31
Aedil joined
06:59
silug left,
silug joined
07:00
sivoais_ left
07:02
sivoais joined
07:03
Pixi left,
ShimmerFairy left
07:35
melezhik joined
08:06
arkiuat left
|
|||
| grondilu | turns out I can do without multithreading: gist.github.com/grondilu/c9882fdb4...7b55781834 | 08:09 | |
| ^this actually works, I can already play against Stockfish. Which is kind of amazing for a short program of 200 lines or so. | 08:10 | ||
| though it does use my Chess modules which are not really short. | |||
|
08:18
arkiuat joined
08:20
abraxxa joined
08:22
arkiuat left
08:23
LainIwakura joined
08:24
Sgeo left,
Sgeo joined
|
|||
| disbot6 | <simon_sibl> I probably have a Cro problem ? github.com/4zv4l/rakubin it used to work great, but just now I realized that when I setup tls and do a request from my browser using the IP instead of the hostname (get a certificate warning), that freezes the whole Cro part (web server), the tcp server part still works somehow | 08:27 | |
|
08:27
arkiuat joined
08:31
abraxxa1 joined
|
|||
| disbot6 | <simon_sibl> so I mean I cannot do any https request after that, but I can still connect to the other port using tcp and create new paste | 08:32 | |
|
08:33
arkiuat left
08:34
abraxxa left
09:03
arkiuat joined
09:08
arkiuat left
09:12
dakkar joined
09:18
Sgeo left
09:38
arkiuat joined
09:43
arkiuat left
09:44
melezhik left
|
|||