!VRULIdgoKmKPzJZzjj:nixos.org

Nix Hackers

902 Members
For people hacking on the Nix package manager itself191 Servers

Load older messages

SenderMessageTime
3 Nov 2021
@ambroisie:belanyi.frAmbroisie
In reply to @tomberek:matrix.org
Thoughts on bundle API? it currently passes a program path, rather than a structured attrset which would better support additional uses.
Are there actual uses of the bundle API? I tried looking for some but only found nix-bundle 		12:43:24
@tomberek:matrix.orgtomberek Ambroisie: I'd like there to be more. (https://github.com/juliosueiras-nix/nix-utils )and i'm putting together a few things ( https://github.com/tomberek/nix-generators ) and noticed this API need. And i'm playing with this PR https://github.com/NixOS/nix/pull/5456 . thought's+opinions+feedback welcome 12:45:47
@tomberek:matrix.orgtomberekI'm thinking of a parallel of nixos-generators, but for any nix closure.12:46:42
@nirahv:matrix.orgnirahv

Are there any thoughts on removing apps.* in flakes since we can rely on heuristics and meta.mainProgram?

I also imagine that one could add something like meta.programs.<name> = "store_path" to make nix run work with multiple output binaries for the same package.

16:11:33
@andi:kack.itandi- Mhm, shouldn't builtins.tryEval (toString {}) return an attrset with the success attribute? Instead it is just barfing at me that it can't convert a set to a string :( 16:13:57
@andi:kack.itandi- It does work with the __toString attribute of a set: builtins.tryEval (toString { __toString = builtins.throw "bar"; }) 16:17:22
@nirahv:matrix.orgnirahv
In reply to @andi:kack.it
Mhm, shouldn't builtins.tryEval (toString {}) return an attrset with the success attribute? Instead it is just barfing at me that it can't convert a set to a string :(

https://nixos.org/manual/nix/unstable/expressions/builtins.html#built-in-functions

"tryEval will only prevent errors created by throw or assert from being thrown. Errors tryEval will not catch are for example those created by abort and type errors generated by builtins."

16:18:39
@andi:kack.itandi- alright, time to duplicate the toString logic then :( 16:19:42
@nirahv:matrix.orgnirahv 
›nix profile remove ''
nix: src/libutil/util.cc:112: nix::Path nix::canonPath(const Path&, bool): Assertion `path != ""' failed.
zsh: abort (core dumped)  nix profile remove ''

oh

18:49:01
@moritz.hedtke:matrix.orgmoritz.hedtke"SUMMARY: AddressSanitizer: 6032 byte(s) leaked in 141 allocation(s)." - it's going down down down (just using GC for everything though)23:14:45
4 Nov 2021
@fzakaria:matrix.orgfzakariaIs there an open issue in Nix that might make a good candidate for a compiler class project?02:45:28
@fzakaria:matrix.orgfzakariaAligning course-work with open-source contributions would sure be swell02:45:48
@fzakaria:matrix.orgfzakariaRun ThreadSanitizer on the Nix codebase ?02:48:39
@tomberek:matrix.orgtomberek fzakaria: hard to find something in Nix itself. But i'm sure hnix or rnix (haskell or rust projects) to do Nix eval. Or go help with Nickel? 02:48:40
@tomberek:matrix.orgtomberekPerhaps do more fuzzing?02:52:54
@abathur:matrix.orgabathur
In reply to @tomberek:matrix.org
Perhaps do more fuzzing?
I think pamplemousse has an ongoing project here, or maybe it's complete already 		02:53:48
@tomberek:matrix.orgtomberekI think the fellowship is over, stopped hearing updates. But yes, that's where I'd start.02:54:41
@tomberek:matrix.orgtomberek * fzakaria: hard to find something in Nix itself. But i'm sure hnix or rnix (haskell or rust projects) would want to do more/better Nix eval. Or go help with Nickel? 02:55:23
@fzakaria:matrix.orgfzakariaCouldn't find the repository they contributed too.02:58:03
@fzakaria:matrix.orgfzakariaI do see the discourse posts though02:58:07
@abathur:matrix.orgabathur tomberek: aside: my ears perked up when you mentioned an updated installer matrix in the rfc106 thread, in part because I've been trying to drum up interest in more of an installer test suite--something that actually confirms Nix installs with different flags/platforms/shells/rcfiles/etc 03:01:05
@abathur:matrix.orgabathur * tomberek: aside: my ears perked up when you mentioned an updated installer matrix in the rfc106 thread, in part because I've been trying to drum up interest in more of an installer test suite--something that actually confirms Nix installs with different flags/platforms/shells/rcfiles/etc (https://discourse.nixos.org/t/installer-test-suite-small-project-s-high-leverage-help-wanted/13662) 03:01:21
@tomberek:matrix.orgtomberekhrm... interesting03:02:51
@tomberek:matrix.orgtomberekI like Graham's approach, it's on VM's so a bit more faithful than relying on Actions.03:07:33
@abathur:matrix.orgabathurI'm not sure what the "right" form is03:15:36
@abathur:matrix.orgabathurbeing able to run them on PRs is great for being able to improve the installer with confidence, but obviously not releasing busted installers is the more important virtue03:17:17
@abathur:matrix.orgabathur for example, the 2.4 single-user installers weren't working on ubuntu because of a change that used echo -e, which isn't compatible with dash, which I guess Ubuntu uses as /bin/sh; actions does test ubuntu, just not the single-user... 03:22:13
@tomberek:matrix.orgtombereklet's do it, pair up sometime?03:23:45
@nirahv:matrix.orgnirahv abathur shellcheck could catch that specific issue 03:35:37
@abathur:matrix.orgabathurit would, but without a test suite someone naively shellchecking the whole installer won't know if they're breaking things that depend on behavior shellcheck warns against03:37:03

There are no newer messages yet.

Back to Room ListRoom Version: 6