17:38 <amirouche> since there is support for OS threads, there is no blockers I can forsee

17:38 <amirouche> ref: https://docs.racket-lang.org/foreign/Operating_System_Threads.html

18:43 <tonyg> Wow, beating ocaml to the punch

18:46 <samth> tonyg: that's really not the same thing

18:47 <samth> ffi/unsafe/os-thread is an unsafe interface to what are effectively pthreads with no support from the runtime for not messing up

18:50 <tonyg> I'd probably say it's not really the same thing, rather than really not the same thing. iirc ocaml's gc was the showstopper for quite a while there; from the lack of gc-related fussing in the docs, it seems like racket-on-chez's gc is already where it needs to be for pthreadlike parallelism

18:51 <tonyg> but also yes, i should be clear that i was mostly joking

18:51 <tonyg> i'm actually really stoked that os-level pthreadlike threads exist.

18:52 <tonyg> weird they've been around since 6.9 but I hadn't noticed! I guess I've been away from Racket for too long

18:52 <samth> tonyg: no, the GC is not at all where it needs to be

18:52 <tonyg> oh.

18:52 <samth> if you use os-threads and both threads GC you get to keep both pieces

18:53 <tonyg> wow!

18:53 <samth> it's unsafe!

18:53 <tonyg> there is... no mention of this on the page

18:53 <samth> it's not remotely a user-facing feature

18:54 <tonyg> i did get that it was unsafe (atomic mode) but not that it was gc-unsafe

18:57 <amirouche> what does GC unsafe means?

18:58 <tonyg> amirouche i'm not using it with any rigor; I mean the idea that gc/memory-safety invariants might end up being broken

19:00 <amirouche> I will look into racket code base, I assume it is used somewhere

19:01 <samth> here's an example: https://github.com/racket/readline/commit/f2c01b705853dd26c51691a1fc78ee743605c8bd

19:02 <samth> it may not be quite as unsafe as I thought

19:06 <samth> here's what the chez docs say about safety: https://cisco.github.io/ChezScheme/csug9.4/threads.html#./threads:s2

19:06 <samth> so I think in fact GC is ok

19:06 <samth> tonyg: also, 6.90 is a pre-7.0 number, not the same as 6.9

19:10 <amirouche> fwiw, my motivation is not only about os-threads; I read the book 'don't teach programming until you read this book', and a publication about how htdp is improving upon sicp; also the availability of a lively package manager , a community, events, comes to mind as stuff that I miss in CS

19:10 <amirouche> also: another adventure!

19:12 <amirouche> First, I will try to contribute to existing projects (that is mostly new for me, I mostly built my own stuff)

19:12 <amirouche> If you have ideas about what issue I can fix, let me know.

19:22 <tonyg> samth: ah good to hear the gc isn't a problem!