14:54 <FromGitter> <sclee15> Hello. Can I ask one question? Suppose I have a fiber that runs expensive blocking function call. Under that situation, are other running fibers can be blocked too? (single core situation)

14:56 <FromGitter> <Blacksmoke16> is the main fiber blocked via a sleep or channel/receive?

14:57 <yxhuvud> sclee15: Yes. You can use Fiber.yield (sometimes, it is somewhat expensive) to give some time to other fibers.

14:59 <yxhuvud> though it depends, not all operations that seems to be blocking is. For example, reading from a socket will block that fiber, potentially for a long time, but will let crystal do other things

15:00 <FromGitter> <sclee15> how can reading from socket allows some cpu time for other thread?

15:00 <FromGitter> <sclee15> i mean other fiber?

15:00 <yxhuvud> it is based on libevent (which in turn is based on a syscall called epoll) that allows that.

15:01 <FromGitter> <sclee15> then I dont have to worry about a expensive sql query in a fiber..

15:02 <yxhuvud> right.

15:04 <FromGitter> <sclee15> oh. right. thank you. @yxhubud and @szabgab

16:46 <riza> what order does crystal read files in a glob-require? is it just whateven Dir[] reads them in?

16:46 <FromGitter> <Blacksmoke16> prob abc

16:47 <riza> what about leading numbers?

16:49 <FromGitter> <Blacksmoke16> do a `ls -l` would prob be that order

16:49 <riza> yeah, that's what I figured it would be :D thanks

16:51 <straight-shoota> Not, it's not just the order that the file system reports

16:51 <straight-shoota> That's unreliable

16:52 <straight-shoota> The paths are actually sorted explicitly: https://github.com/crystal-lang/crystal/blob/5410318bc693629bf50ecf97abee2425f14c1f04/src/compiler/crystal/crystal_path.cr#L189-L190

16:59 <riza> hhhhhhhmmmmmmm

16:59 <FromGitter> <Blacksmoke16> fwiw id advise against using glob requires

17:00 <FromGitter> <Blacksmoke16> esp when files may depend on files earlier/later

17:01 <riza> does that sort take into account numbers, ro is it sorting based on (essentially) ascii table

18:55 <FromGitter> <jrei:matrix.org> I agree 💯 with @beta-ziliani about thr errors. We have type safety, but miss error safety now :/

18:56 <FromGitter> <jrei:matrix.org> I did experiment a bit in the past https://github.com/j8r/error.cr

19:00 <FromGitter> <jrei:matrix.org> Using exceptions are convenient, like not having to deal with types. But it can backslash (and will) at runtime

19:10 <FromGitter> <MrSorcus> https://github.com/uuid6/prototypes/ - who want create prototype in Crystal? 😸

19:16 <FromGitter> <jrei:matrix.org> There are also UUIDv7 and UUIDv8

19:18 <FromGitter> <suskeyhose:matrix.org> is that you johnny, or just someone else who loves kung fury just a little too much? :P

19:20 <FromGitter> <MrSorcus> @jrei:matrix.org yeap.

19:51 <FromGitter> <jrei:matrix.org> srasu (https://matrix.to/#/@suskeyhose:matrix.org) not johnny haha. Good catch, that's Hackerman. I like the movie, and some people said I was like Hackerman - hence the image

20:18 <FromGitter> <suskeyhose:matrix.org> Ah, fair. Johnny is a contributor to one of my repos and he always has various hackerman avatars