| 15 Sep 2021 |
Las | TBF, Nix is supposed to be on a faster schedule now | 15:24:44 |
baloo | * adding patches to nixpkgs is the only leverage I have :( | 15:25:12 |
baloo | I'll be more than happy to drop those patches in favor of a release | 15:25:45 |
eyJhb | Unsure how this command works - https://github.com/NixOS/nix/blob/master/doc/manual/local.mk#L69, where it calls __dump-builtins, what does it actually do? | 16:09:06 |
eyJhb | Ie. what command would I run to do the same? | 16:09:41 |
Jan Tojnar | eyJhb: it is a subcommand of nix command: nix __dump-builtins | 17:26:57 |
eyJhb | In reply to @jtojnar:matrix.org eyJhb: it is a subcommand of nix command: nix __dump-builtins i just get unknown command when I do it, or rather that it isn't a known subcommand :/ | 17:27:53 |
Jan Tojnar | do you have unstable nix? | 17:28:11 |
eyJhb | Just use the one in my unstable nixpkgs, so the default there. Should I clone nix and do nix-build and use that? | 17:29:22 |
Jan Tojnar | nixUnstable attribute works | 17:30:05 |
Jan Tojnar | * the one from nixUnstable attribute works | 17:30:13 |
eyJhb | I Will try that tomorrow, thanks! :) | 18:05:19 |
| 16 Sep 2021 |
| Rev. CornWallace III (novus ordo seclorum) changed their display name from Corn Wallet to CornWallace. | 01:56:27 |
| 17 Sep 2021 |
| mkg20001 joined the room. | 00:10:47 |
| 19 Sep 2021 |
Las | Does anyone know if any of the builtins.unsafe* functions are actually unsafe? | 16:00:39 |
Las | Can you misuse them such that some property of Nix is broken? | 16:00:51 |
Alyssa Ross | I believe so | 16:02:06 |
Alyssa Ross | for example, with unsafeDiscardStringContext you can break the dependency calculation | 16:02:30 |
Las | But does that result in anything other than a borked package? | 16:05:01 |
Las | * But does that result in anything other than a borked output/derivation? | 16:05:06 |
Las | * But can that result in anything other than a borked output/derivation? | 16:05:12 |
Las | Could it e.g. be unsafe from a security perspective? | 16:05:39 |
Alyssa Ross | I don't think so, but nobody said unsafe meant insecure | 16:25:06 |
Las | Generally in Haskell and Rust, unsafe means that it can break the system. | 16:51:41 |
Las | i.e. you have no guarantees about the behavior if you pass in something wrong | 16:51:57 |
Las | * i.e. you have no guarantees about the behavior if you pass in something wrong. | 16:51:59 |
tomberek | Is the latest rendered docs (on master after 49a932fb18add471feefc469fb45fc44313bd5c6) hard to read for anyone else? Lowdown seems to try to highlight things, but most terminals I've tried this in make it more difficult to read. | 17:25:20 |
tomberek | It is rendering the markdown with these codes:
printf "\u001b[1;47;92m Testing \u001b[0m" | 17:44:20 |
sterni | In reply to @Las:matrix.org i.e. you have no guarantees about the behavior if you pass in something wrong. well it is a guarantee nix gives you usually that every referenced store path in a derivation output also has to exist for example | 18:55:01 |
Las | Is that really guaranteed? | 18:55:58 |