nimforum mirror - Malebogia

10198 :: Malebogia

[2023-05-15T10:00:02+02:00]

View Orginal

Araq (orginal) [2023-05-15T10:00:02+02:00] view original

Experiments with thread pools: https://github.com/Araq/malebolgia

Contributions/feedback welcome.

Why now? Because it's a great new pun and I want to take credit for it.

moigagoo (orginal) [2023-05-15T10:04:14+02:00] view original

Why now? Because it's a great new pun and I want to take credit for it.

Ahh, good old naming-driven development. I'm a fan myself :-)

Araq (orginal) [2023-05-15T11:09:50+02:00] view original

Ahh, good old naming-driven development

Nice, I've never heard this term before.

moigagoo (orginal) [2023-05-15T11:55:49+02:00] view original

Well, it's a term I invented some time ago to describe my own habit of making up product names with puns and then trying to invent the product for the name :-)

federico3 (orginal) [2023-05-15T12:49:27+02:00] view original

Looks similar to https://vorpus.org/blog/notes-on-structured-concurrency-or-go-statement-considered-harmful/ - which is quite promising. I wonder if something similar could be used with async workers to solve the problem of unhandled exceptions.

BTW is the name meant to refer to https://en.wikipedia.org/wiki/Malebolgia or https://en.wikipedia.org/wiki/Malebolge ?

jasonfi (orginal) [2023-05-15T13:58:47+02:00] view original

Is it feasible to create channels to send to, or receive from the new thread, but in an easy way?

Araq (orginal) [2023-05-15T15:20:30+02:00] view original

Looks similar to https://vorpus.org/blog/notes-on-structured-concurrency-or-go-statement-considered-harmful/ - which is quite promising.

That was one source of inspiration, yes.

BTW is the name meant to refer to https://en.wikipedia.org/wiki/Malebolgia or https://en.wikipedia.org/wiki/Malebolge ?

To https://en.wikipedia.org/wiki/Malebolgia

demetera (orginal) [2023-05-15T16:32:30+02:00] view original

Finally long-awaited thread abortion proc without using channels explicitly. Awesome package. This should be a part of stdlib at some point.

termer (orginal) [2023-05-15T17:28:47+02:00] view original

It's referring to "Male Bulge", an ancient concept common not only philosophers, but also software developers, engineers, and other male-dominated industries.

juancarlospaco (orginal) [2023-05-15T20:28:52+02:00] view original

🔜 Malbolge backend ? https://en.wikipedia.org/wiki/Malbolge

Araq (orginal) [2023-08-30T23:04:16+02:00] view original

Done.

blackmius (orginal) [2023-08-31T08:14:26+02:00] view original

the new 3 procs has the same blocksize iterator pattern, wouldn't it better to extract it and reuse?

Araq (orginal) [2023-08-31T08:25:43+02:00] view original

I think you cannot really implement reduce and map with the same efficiency but be my guest and prove me wrong.

imagio (orginal) [2023-12-06T16:14:53+01:00] view original

Interesting stuff! I just discovered Nim and was poking around to see if there were any "structured concurrency" implementations when I found this. If you're not familiar with them already you might want to take a look at Effect in TS https://effect.website/ and ZIO in Scala https://github.com/zio/zio for inspiration. They provide an (IMO) absolutely beautiful model for writing fast, concurrent, safe, fault tolerant programs.

matkuki (orginal) [2024-11-25T09:30:16+01:00] view original

Can someone look at this Malobolgia issue: https://github.com/Araq/malebolgia/issues/37

Thanks

matkuki (orginal) [2024-12-11T15:09:19+01:00] view original

I'm using this in a procedure:


proc sound_alarm() =
  const filename = "mixkit-uplifting-flute-notification-2317.wav"
  malebolgia_master.spawn soundutils.play_wav(filename) -> malebolgia_result

Both malebolgia_master and malebolgia_result are globals. But the spawn always blocks until the sound is played to the end. Is there a way to spawn-and-forget?

matkuki (orginal) [2024-12-15T14:53:27+01:00] view original

Sorry, if you set the internal constant ThreadPoolSize to 1, then it always block, setting it to higher works as expected ... but there is a thread disposal problem if that value is greater than 1 :)

planetis (orginal) [2024-12-15T16:51:02+01:00] view original

What do you mean there's a "thread disposal" problem. Threads in the threadpool remain active until the program terminates or you explicitly call panicStop.

matkuki (orginal) [2024-12-15T17:20:52+01:00] view original

panicStop hangs forever in at the joinThreads(thr) line, if you at the source code. This is on Windows.

Mirror of forum.nim-lang.org

10198 :: Malebogia