yilin.wei10
2020-12-19 13:57:18

@samth I worked around the issue in the end using prop:procedure and prop:struct-info .


popa.bogdanp
2020-12-19 18:03:10

Does an unsafe-poller need to do anything specific to cooperate with sync/enable-break? Since its poll argument is run in atomic mode, I would assume the answer is “no”, but I’m seemingly able to race a call to (sync/enable-break my-evt) against a break and end up with both the break occurring and the event consumed. This only happens on CS, never on BC. Context <https://github.com/racket/unix-socket/pull/4/files#r544170166|here. >


mflatt
2020-12-19 18:04:50

That’s almost certainly a CS bug.


popa.bogdanp
2020-12-19 18:07:25

Should I file an issue? I’m running commit 627c45e8d4, if that’s relevant.


mflatt
2020-12-19 18:08:01

Yes, please


gierczak.o
2020-12-19 20:12:05

@gierczak.o has joined the channel


nurrukia1
2020-12-19 20:25:06

@nurrukia1 has joined the channel


notjack
2020-12-19 23:39:47

the efficiency advantages of mutable strings could all be achieved with mutable string builders instead


kellysmith12.21
2020-12-19 23:45:24

I may need to look into making a string builder type, then.


samth
2020-12-19 23:48:37

open-output-string is a string builder, although it doesn’t always have the most convenient API


samth
2020-12-19 23:49:56

However, I’m skeptical that all efficiencies are possible just with string builders