|
Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes. Set by lizmat on 24 May 2021. |
|||
|
01:48
MasterDuke joined
07:10
sena_kun joined
08:30
sena_kun left
11:54
MasterDuke left
|
|||
| Woodi | are problems with $/ only in concurent code ? but in concurent code CWD can change too, isn't this similiar situation ? | 14:46 | |
| timo | moarvm doesn't use the regular process-wide CWD, we have virtual CWD that we set whenever we fork/exec | 14:51 | |
| Woodi | fork/exec is process too so how virtual work ? | 14:53 | |
| but re $/: cwd is just example. any other magical var can have same problem with many threads changing them. but question is: do $/ have problems only in concurent code ? | 14:54 | ||
| timo | the problems we've seen was when right in the middle of code that we consider linear the value (and type) inside the variable gets changed. you can't do that without concurrency | 15:03 | |
| i also count signal handlers in concurrency here | |||
| docs.libuv.org/en/v1.x/process.html we use this to spawn processes, it takes cwd as a string and takes care of the rest | 15:06 | ||
| Woodi | what I mean is: it can be that $/ work ok but concurent code changes program state ($/ in this case) that is somewhat in ~upper program logic like cwd (global) or multithreaded array sum (local to block). just saying becouse plans to change spec already was considered :) | 16:10 | |
| timo | $/ is a lexical variable | 16:46 | |
|
18:12
sena_kun joined
21:34
sena_kun left
21:37
sena_kun joined
21:42
sena_kun left
22:42
MasterDuke joined
|
|||