!VRULIdgoKmKPzJZzjj:nixos.org

Nix Hackers

926 Members
For people hacking on the Nix package manager itself194 Servers

Load older messages


SenderMessageTime
30 Jan 2026
@faukah:faukah.comfaukahimage.png
Download image.png
13:37:52
31 Jan 2026
@eveeifyeve:matrix.orgeveeifyevePlease respond that you have reproduced it and describe information so we can reproduce it and 👍️ (if you think it's important).17:22:21
@eveeifyeve:matrix.orgeveeifyeve* Please respond to this issue you have linked and say you have reproduced it with describe information so we can reproduce it and 👍️ (if you think it's important).17:22:43
@eveeifyeve:matrix.orgeveeifyeve* Please respond to this issue you have linked and say you have reproduced it with describe information so we can reproduce it while also giving it a 👍️ (if you think it's important).17:22:55
@eveeifyeve:matrix.orgeveeifyeveDoes anyone know how to run tests just by them selves without meson. Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=1504117:23:45
@eveeifyeve:matrix.orgeveeifyeve* Does anyone know how to run tests just by them selves without meson. Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=15041 to fix them17:23:50
@eveeifyeve:matrix.orgeveeifyeve* Does anyone know how to run tests just by them selves without meson?? Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=15041 to fix them17:24:22
@eveeifyeve:matrix.orgeveeifyeve * Does anyone know how to run tests just by them selves without meson?? Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=15041 to fix them and the docs example doesn't work while being in a nix develop. 17:24:56
@eveeifyeve:matrix.orgeveeifyeve* Does anyone know how to run tests just by them selves without meson?? Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=15041 to fix them and the docs example doesn't work.17:25:21
@eveeifyeve:matrix.orgeveeifyeve It should be this: TEST_NAME="post-hook" NIX_REMOTE='' PS4='+(${BASH_SOURCE[0]-$0}:$LINENO) '' tests/functional/post-hook.sh but it's not running the thing... 17:25:47
@eveeifyeve:matrix.orgeveeifyeve Nvm it's TEST_NAME="post-hook" NIX_REMOTE='' PS4='+(${BASH_SOURCE[0]-$0}:$LINENO)' tests/functional/post-hook.sh. 17:27:43
@eveeifyeve:matrix.orgeveeifyeveReported: https://github.com/NixOS/nix/issues/1512617:32:20
@eveeifyeve:matrix.orgeveeifyeve John Ericson: Hmm interesting error: post-build-hook: error: not an absolute path: 'netrc'. Idk where to start to fix that. 17:47:04
@eveeifyeve:matrix.orgeveeifyeve

Look at this, I am thinking it something caused by:
⁨⁨```cpp
: nixConfDir(canonPath(getEnvOsNonEmpty(OS_STR("NIX_CONF_DIR"))
.transform([](auto && s) { return std::filesystem::path(s); })
.value_or(resolveNixConfDir())))

17:59:38
@eveeifyeve:matrix.orgeveeifyeve *

Look at this, I am thinking it something caused by:

⁨⁨```cpp
: nixConfDir(canonPath(getEnvOsNonEmpty(OS_STR("NIX_CONF_DIR"))
.transform([](auto && s) { return std::filesystem::path(s); })
.value_or(resolveNixConfDir())))

17:59:51
@eveeifyeve:matrix.orgeveeifyeve *

Look at this, I am thinking it something caused by:

⁨⁨```
nixConfDir(canonPath(getEnvOsNonEmpty(OS_STR("NIX_CONF_DIR"))
.transform([](auto && s) { return std::filesystem::path(s); })
.value_or(resolveNixConfDir())))

18:00:08
@eveeifyeve:matrix.orgeveeifyeve *

Look at this, I am thinking it something caused by:

⁨⁨cpp nixConfDir(canonPath(getEnvOsNonEmpty(OS\_STR("NIX\_CONF\_DIR")) .transform(\[\](auto && s) { return std::filesystem::path(s); }) .value\_or(resolveNixConfDir())))

18:00:19
@eveeifyeve:matrix.orgeveeifyeve *

Look at this, I am thinking it something caused by:

⁨⁨nixConfDir(canonPath(getEnvOsNonEmpty(OS\_STR("NIX\_CONF\_DIR")) .transform(\[\](auto && s) { return std::filesystem::path(s); }) .value\_or(resolveNixConfDir())))

18:00:25
@eveeifyeve:matrix.orgeveeifyeve * Look at this, I am thinking it something caused by:
⁨⁨nixConfDir(canonPath(getEnvOsNonEmpty(OS\_STR("NIX\_CONF\_DIR")) .transform(\[\](auto && s) { return std::filesystem::path(s); }) .value\_or(resolveNixConfDir())))
18:00:29
@eveeifyeve:matrix.orgeveeifyeve *

Look at this, I am thinking it something caused by:
⁨⁨```cpp
: nixConfDir(canonPath(getEnvOsNonEmpty(OS_STR("NIX_CONF_DIR"))
.transform([](auto && s) { return std::filesystem::path(s); })
.value_or(resolveNixConfDir())))

18:01:32
@eveeifyeve:matrix.orgeveeifyeve * Look at this, I am thinking it something caused by:
⁨⁨ : nixConfDir(canonPath(getEnvOsNonEmpty(OS\_STR("NIX\_CONF\_DIR")) .transform(\[\](auto && s) { return std::filesystem::path(s); }) .value\_or(resolveNixConfDir())))
18:12:17
@eveeifyeve:matrix.orgeveeifyeve * Look at this, I am thinking it something caused by:
⁨⁨: nixConfDir(canonPath(getEnvOsNonEmpty(OS\_STR("NIX\_CONF\_DIR")) .transform(\[\](auto && s) { return std::filesystem::path(s); }) .value\_or(resolveNixConfDir())))
18:12:23
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)
In reply to @eveeifyeve:matrix.org
Does anyone know how to run tests just by them selves without meson?? Because I am trying to run tests to get to the point of failing tests both locally and CI of this pr: https://github.com/NixOS/nix/actions/runs/21536152140/job/62062248779?pr=15041 to fix them and the docs example doesn't work.
In my experience that’s a moot point. Just run via meson test —interactive
18:50:29
@figsoda:matrix.orgfigsoda joined the room.20:22:49
@figsoda:matrix.orgfigsoda Is there some way to specify the name of the output path when using nix flake prefetch or fetchTree? I digged a little through the source code, and it seems like name is an attribute that defaults to source and passed around, but I couldn't find any way to change it to something other than the default. 20:27:47
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)
In reply to @figsoda:matrix.org
Is there some way to specify the name of the output path when using nix flake prefetch or fetchTree? I digged a little through the source code, and it seems like name is an attribute that defaults to source and passed around, but I couldn't find any way to change it to something other than the default.
Don’t think so, it’s only allowed for fetchGit
20:44:05
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)FetchTreeParams::allowNameArgument20:44:24
@figsoda:matrix.orgfigsodaThanks for the pointers, do you know if there is a design reason why that's the case? Or is it just that they haven't got around to implementing it, and it was just there so fetchGit can use fetchTree20:48:11
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)
In reply to @figsoda:matrix.org
Thanks for the pointers, do you know if there is a design reason why that's the case? Or is it just that they haven't got around to implementing it, and it was just there so fetchGit can use fetchTree
Honestly I don’t know. “source” is special-cases in all sorts of places in the store layer and the justification is probably along the lines of why that used to be preferred for FOD fetchers in nixpkgs - when changing the fetcher the outpath would stay the same. Whether that’s a good idea is another question
20:50:09
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)
In reply to @figsoda:matrix.org
Thanks for the pointers, do you know if there is a design reason why that's the case? Or is it just that they haven't got around to implementing it, and it was just there so fetchGit can use fetchTree
* Honestly I don’t know. “source” is special-cased in all sorts of places in the store layer and the justification is probably along the lines of why that used to be preferred for FOD fetchers in nixpkgs - when changing the fetcher the outpath would stay the same. Whether that’s a good idea is another question
20:50:51

Show newer messages


Back to Room ListRoom Version: 6