!lymvtcwDJ7ZA9Npq:lix.systems

Lix Development

422 Members
(Technical) development of Lix, the package manager, a Nix implementation. Please be mindful of ongoing technical conversations in this channel.142 Servers

Load older messages


SenderMessageTime
18 May 2024
@ckie:ckie.devmei 🌒&uh i'm not sure, just OOM'd after writing that msg20:01:46
@ckie:ckie.devmei 🌒& luckily i made it recover nicely so im back here already. :p mapAttrsRecursiveCond (as: true) id nodes.cookiemonster.options 20:02:18
@puck:puck.moepuckah, that is in fact infinitely recursive non-cycle20:02:28
@ckie:ckie.devmei 🌒&eats all the memory and a few cpu cores and yeah20:02:31
@ckie:ckie.devmei 🌒&not good20:02:36
@ckie:ckie.devmei 🌒&
In reply to @puck:puck.moe
ah, that is in fact infinitely recursive non-cycle
bad ux,,
20:02:45
@ckie:ckie.devmei 🌒&lost my big repl thing cuz i printed a recursive thing20:03:08
@puck:puck.moepuckyeahhhhh we should have an issue for repl history imploding20:03:21
@ckie:ckie.devmei 🌒&and recurse limit 🥺20:03:32
@ckie:ckie.devmei 🌒&maybe even configurable interactively-only in the repl20:03:47
@ckie:ckie.devmei 🌒&silly code should not hang the machine20:04:18
@ckie:ckie.devmei 🌒&ill make the issues20:05:01
@ckie:ckie.devmei 🌒& REPL history should be saved immediately #328 20:07:35
@raitobezarius:matrix.orgraitobezariusif i remember, there was a crude recurse limit detector20:07:56
@raitobezarius:matrix.orgraitobezariusbut i think eelco wanted to remove it?20:08:01
@puck:puck.moepucki think we added one in, yeah20:08:37
@puck:puck.moepuckbut that doesn't help wider recursion issues, which i think this was20:08:47
@ckie:ckie.devmei 🌒&not making the other ticket since idk enough20:10:46
@ckie:ckie.devmei 🌒&or i could make a "this expr blows repl up" thing20:11:00
@raitobezarius:matrix.orgraitobezariusfwiw, tvix has no recursion limit20:12:21
@raitobezarius:matrix.orgraitobezarius(constant stacksize recursion IIRC)20:12:41
@raitobezarius:matrix.orgraitobezariuseffectively, this is obviously not a solution for your case20:12:52
@ckie:ckie.devmei 🌒&someday they will make a wasm engine in nix and you will play doom over the filesystem and __trace20:12:57
@raitobezarius:matrix.orgraitobezariusI think L-as was interested into that in CppNix20:13:12
@ckie:ckie.devmei 🌒&someone was with the wait+readFile thing yuh20:13:26
@ckie:ckie.devmei 🌒&but theres no synchronization primitive 20:13:34
@ckie:ckie.devmei 🌒&tho.. if you run fast enough maybe you could just loop over a readFile 20:14:08
@grimmauld:grimmauld.deGrimmauld (moving to @grimmauld:grapevine.grimmauld.de)
In reply to @ckie:ckie.dev
someday they will make a wasm engine in nix and you will play doom over the filesystem and __trace
i mean, you could technically do it the nixos way running during some phase of a "package" i think?though thats not really nix, so kinda cheating
20:23:01
@ckie:ckie.devmei 🌒&it'd still b cool20:23:17
19 May 2024
@0x4d6165:catgirl.cloud0x4d6165 joined the room.05:25:11

Show newer messages


Back to Room ListRoom Version: 10