🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
Geth rakudo: MasterDuke17++ created pull request #5483:
Attempt at adding debugging to delete-by-compiler to test git…
02:19 MasterDuke joined 09:35 sena_kun joined 14:00 MasterDuke left
Geth rakudo: vrurg++ created pull request #5484:
Move close to complete generic instantiation, iteration 2
17:18 gfldex joined
ugexe m: say "aliwjefqjweoifjqwojf".IO.rmdir 21:33
camelia Failed to remove the directory '/home/camelia/aliwjefqjweoifjqwojf': Failed to rmdir: no such file or directory
in block <unit> at <tmp> line 1
ugexe m: say rmdir("aliwjefqjweoifjqwojf");
camelia []
ugexe that doesn't seem great 22:09
probably the same issue with unlink and other IO subs that take multiple paths 22:10
vrurg ugexe: BTW, after giving some more thinking to the problem with stashes, I think parallel compiling could be suffering from it too. Moreover, it would be a flapper then because it depends on the order of loading too. For example, another workaround for the Path::Finder test would be to move 'use File::Temp' down a couple of line, so that the test helper module gets loaded first. 22:58
23:13 sena_kun left, sena_kun joined 23:14 sena_kun left
ugexe I think the issue with parallel module loading is two things trying to read/write the stash at the same time. So there isn't exactly any confusion involved in parallel module loading issue, but rather something similar to trying to read/write to a Hash from multiple threads 23:29
vrurg And this brings me back to the dream of having a debug mode where unsafe concurrent hash access would throw with backtraces available for every thread involved. That'd make it possible to pinpoint so many races! 23:52