00:53 Aedil joined 01:16 hulk joined, kylese left 01:29 sibl joined 01:53 sibl left 02:01 Sgeo_ left, Sgeo joined 02:15 hulk left, kylese joined 02:21 arkiuat left 02:28 arkiuat joined 02:38 stanrifkin left 02:39 stanrifkin joined 02:51 hulk joined 02:52 kylese left
aruniecrisps @SmokeMachine Wanted to see whether I could make a PR to amend the README for Red to also indicate that you may need to add -exclude="pq:ver<5>:from<native>" when installing Red 03:14
03:53 stanrifkin_ joined 03:55 stanrifkin left 04:22 Aedil left 04:28 stanrifkin joined 04:30 stanrifkin_ left 04:52 Aedil joined 05:13 arkiuat left 05:15 arkiuat joined 05:21 arkiuat left 05:34 arkiuat joined 05:41 arkiuat left 05:53 arkiuat joined 06:01 camelia joined 06:02 arkiuat left, LainIwakura left 06:15 melezhik joined
melezhik . 06:15
06:15 arkiuat joined 06:19 arkiuat left 06:48 arkiuat joined 06:54 arkiuat left 06:56 arkiuat joined 07:00 arkiuat left 07:12 arkiuat joined 07:17 arkiuat left
SmokeMachine aruniecrisps: sure, please do it! 07:21
07:30 arkiuat joined 07:32 Aedil left 07:56 arkiuat left 07:57 arkiuat joined 08:01 arkiuat left 08:13 arkiuat joined 08:22 arkiuat left, melezhik left 08:23 LainIwakura joined 08:29 mimosa joined 08:32 ACfromTX left 08:49 arkiuat joined 08:54 arkiuat left 09:01 Sgeo left 09:09 arkiuat joined 09:13 arkiuat left 09:16 lichtkind_ joined 09:23 melezhik joined
melezhik . 09:23
If someone wish to participate in brownie demo - here is information - irclogs.raku.org/raku-dev/2025-10-24.html#09:05 , thanks 🙏 09:24
09:27 sftp joined, sftp left, sftp joined 09:42 arkiuat joined 09:47 arkiuat left 09:57 Guest15 joined, Guest15 left 10:15 arkiuat joined 10:19 arkiuat left 10:47 arkiuat joined 10:52 arkiuat left 11:17 arkiuat joined 11:24 arkiuat left 11:43 arkiuat joined 11:48 arkiuat left
melezhik . 11:58
12:01 arkiuat joined 12:06 arkiuat left 12:20 arkiuat joined 12:26 arkiuat left 13:00 arkiuat joined 13:10 apa_c joined 13:15 melezhik_ joined 13:26 melezhik_ left 13:28 melezhik_ joined 13:56 melezhik_ left 14:01 melezhik_ joined 14:06 melezhik_ left 14:19 apa_c left 14:21 apa_c joined 14:38 ACfromTX joined 14:44 apa_c left
.landyacht. I'm trying to design a threadpool that handles a mix of future-scheduled and ASAP jobs. My current thought is to have two queues, one for work ready to be performed which the worker threads read from, and one for future jobs that a coordinator thread reads from and moves them to the former queue at the appropriate time. My idea is, if the soonest scheduled job say, n seconds away, to have the coordinator queue sleep n seconds 15:03
and then do its thing. But I need some way to "wake" that thread if another future-scheduled job comes in, say, n - 1 seconds away...
the coordinator thread*
is there a mechanism in Raku to sleep a thread but be able to wake it sooner? 15:04
or will I just need to decide on a time resolution at which the coordinator thread will check regardless of when the soonest job is scheduled 15:05
15:05 Aedil joined
rcmlz Is that perhaps something than can use - instead of sleep - something like this: rosettacode.org/wiki/IPC_via_named_pipe#Raku 15:16
15:24 Romanson joined
.landyacht. react/whenever might well come in handy.. not sure, might need to rework my approach entirely 15:24
15:40 LainIwakura left 15:58 vasko453 left, vasko4535 joined 16:02 ds7832 joined, melezhik_ joined 16:05 melezhik_ left 16:17 LainIwakura joined 16:19 melezhik_ joined 16:21 LainIwakura left 16:25 LainIwakura joined 16:28 melezhik_ left, melezhik_ joined 16:31 ds7832 left 16:33 melezhik left 16:34 LainIwakura left 16:39 LainIwakura joined 16:40 apa_c joined 16:44 melezhik joined 16:46 melezhik_ left 16:59 ds7832 joined
rcmlz using „bussy wait“-approaches feels so 90‘s 😀 17:04
Perhaps the whole thing can be designed entirely without file system (and thus OS specific) events - just is case you aim for Windows & Linux users … 17:07
and still not use bussy waiting 17:08
17:41 hulk left, kylese joined 17:44 apa_c left 17:52 melezhik_ joined, melezhik_ left 17:58 arkiuat left 18:09 arkiuat joined
m_zero Have the coordinator thread wait on a semaphore - when it triggers, it evaluates the current queue, moves jobs to be done to the ready queue, and launches a thread that: wakes in n seconds (the soonest job in the waiting queue) then signals the semaphore. If when some other code puts something on the waiting queueu - just signal the semaphore and let the coordinator sort it out. The idea here is that waking the coordinator is 18:10
always safe - it just looks at the state of the queues and adjusts. If it ends up setting off extra timers (like if something comes in that is to be done sooner than the previous next task) - that's okay.
18:11 avuserow_ is now known as avuserow
.landyacht. Ahhh that’s a very clever solution, I think that will work for what I want. Thanks! 18:13
18:15 arkiuat left 18:26 arkiuat joined
Voldenet busy wait solution is actually used for spinlocks though 18:35
it is a good solution if lock/monitor isn't held for very long
18:35 Romanson left
Voldenet if you simply want it to work, plain locks are good enough, you might get a bit clever and use Lock.condition (which I think uses uv_cond_wait underneath, but I'm not sure) 18:37
it removes need for busy loop 18:38
18:49 Sgeo joined
Voldenet rough example of basic thread pool 0x0.st/K2VR.raku 18:50
19:04 corwin is now known as pii
Voldenet with some random waits you can see that it's more or less fair 0x0.st/K2Wz.raku 19:05
19:05 apa_c joined
Voldenet (as they should, since only idle tasks are waiting for items) 19:05
I'm deliberately not using .join or more primitives to highlight the conditional variables, which are incredibly useful for this specific case 19:06
19:19 Guest70 joined 19:33 melezhik left 19:35 pii is now known as corwin, corwin is now known as pii
Guest70 When using map, is there a way to get the index? 19:37
For example, in Javascript, the map callback gets up to three arguments: the data element, its index within the sequence, and the full sequence. 19:41
Failing that, is there a way to define a map-like operator that does get the index?
mc2 Guest70: < a b c >.kv.map: { say "$^k = $^v" } 19:43
Guest70: < a b c >.pairs.map: { say "{.key} = {.value}" } 19:44
Guest70 Thanks 19:45
mc2 you're welcome 19:46
19:47 LainIwakura left 19:55 Geth left 20:45 librasteve_ joined 21:00 apa_c left 21:02 mimosa left 21:10 kylese left 21:44 Aedil left 22:16 arkiuat left 22:28 arkiuat joined 22:35 arkiuat left 22:46 lichtkind_ left 22:49 tailgate left 22:50 tailgate joined 22:52 ds7832 left, arkiuat joined 22:57 arkiuat left 23:09 arkiuat joined 23:11 Geth joined 23:16 arkiuat left 23:21 arkiuat joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!