qman: init at 1.5.1

[yzhou216]

Things done

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• Tested, as applicable:
  • NixOS tests in nixos/tests.
  • Package tests at passthru.tests.
  • Tests in lib/tests or pkgs/test for functions and "core" functionality.
• Ran nixpkgs-review on this PR. See nixpkgs-review usage.
• Tested basic functionality of all binary files, usually in ./result/bin/.
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other READMEs.

---

Add a 👍 reaction to pull requests you find important.

[AaronVerDow]

I had to add documentation.man.generateCaches = true; to my configuration.nix for this to work, but I don't think that's within the scope of an individual package.

Added this to my personal configs, looks good and works well.

[yzhou216]

I had to add documentation.man.generateCaches = true; to my configuration.nix for this to work, but I don't think that's within the scope of an individual package.

Added this to my personal configs, looks good and works well.

Interesting, I didn't have to enable that. What do you mean by not working?

[AaronVerDow]

I had to add documentation.man.generateCaches = true; to my configuration.nix for this to work, but I don't think that's within the scope of an individual package.
Added this to my personal configs, looks good and works well.

Interesting, I didn't have to enable that. What do you mean by not working?

I get the error:

'apropos .' failed; did you run mandb/makewhatis?

[yzhou216]

I had to add documentation.man.generateCaches = true; to my configuration.nix for this to work, but I don't think that's within the scope of an individual package.
Added this to my personal configs, looks good and works well.

Interesting, I didn't have to enable that. What do you mean by not working?

I get the error:

'apropos .' failed; did you run mandb/makewhatis?

Is documentation.man.generateCaches = true; slow for you? It takes forever on my machine, maybe something is blocking.

[yzhou216]

@kpbaks Does this look okay to you? I also had issues with versionCheckHook, any ideal on how to work around it?

[kpbaks]

@kpbaks Does this look okay to you? I also had issues with versionCheckHook, any ideal on how to work around it?

Yes LGTM. I have not come up with a way to solve the issue with versionCheckHook. My only hunch is that the version() function in the qman codebase outputs a string array of wchar_t and not char. But when I run

[[ $(qman --version) =~ "1.5.0" ]] && echo "matches"

it matches. Let's just leave it out for now.

[AaronVerDow]

I had to add documentation.man.generateCaches = true; to my configuration.nix for this to work, but I don't think that's within the scope of an individual package.
Added this to my personal configs, looks good and works well.

Interesting, I didn't have to enable that. What do you mean by not working?

I get the error:

'apropos .' failed; did you run mandb/makewhatis?

Is documentation.man.generateCaches = true; slow for you? It takes forever on my machine, maybe something is blocking.

Not that I had noticed. I'm not sure how qman works without that option, on NixOS at least. Nixpkgs on another OS will probably be fine.

[yzhou216]

[[ $(qman --version) =~ "1.5.0" ]] && echo "matches"

Sounds good, I'm going to spend some time and see if I can reproduce this problem, if so I'll open an issue with regard to versionCheckHook.

[kpbaks]

I think we can merge it as it is now, and then in the future extend it with a versionCheckHook once the issue with it has been resolved.

[lovelymono]

Can be updated to 1.5.1 while we're at it.

[lilyball]

I ended up submitting essentially a duplicate of this at #496407, but I've got some tweaks you don't have. I'm working on tweaking my derivation to account for the differences between yours and mine, I'll have that ready to propose soon.

[lilyball]

I've also figured out the versionCheckHook problem, it looks like qman crashes if the TERM env var isn't present.

[lilyball]

I just updated #496407 to account for the differences between my original version and this PR. My updated version is similar to this PR except it uses versionCheckHook, it properly spits out the test dependencies, it removes the installed sample config file (because it would be non-editable and so its presence breaks the ability to use /etc/xdg/qman/qman.conf and also changes some of the settings away from the documented defaults), it uses /usr/bin/open on darwin instead of using xdg-utils, and it also puts the correct paths into the manpage instead of claiming to be using /run/current-system/sw/bin. I skipped editing doc/TROUBLESHOOTING.md as the /usr/bin references there are just example configuration values. I also skipped setting propagatedUserEnvPkgs as this PR does since the default config values are absolute paths rather than relying on PATH. I also added the two maintainers from this PR into my version.

If you like my changes, it doesn't matter to me whether you copy them into this PR or close it in favor of my PR (though if you copy it into this one you should remember to update the PR title for the version update). If you disagree with any of the changes just let me know.

[yzhou216]

I ended up submitting essentially a duplicate of this at #496407, but I've got some tweaks you don't have. I'm working on tweaking my derivation to account for the differences between yours and mine, I'll have that ready to propose soon.

Great work!

[lilyball]

I resubmitted my version of this as a PR against your branch as requested: yzhou216#2. BTW don't forget to fix the title of this PR!

[yzhou216]

@lilyball Is propagatedUserEnvPkgs still needed for this?

[yzhou216]

nixpkgs-review result

Generated using nixpkgs-review-gha

Command: nixpkgs-review pr 434438
Commit: 57ff75382665ef4dbf5f580e6c519ffdd7617145 (subsequent changes)
Merge: f003c6a86b0a8cbee6c664fe686b7fb1f26ba600

Logs: https://github.com/yzhou216/nixpkgs-review-gha/actions/runs/22924867279

Download packages from cache:

• x86_64-linux

  nix-store -r --add-root nixpkgs-pr-434438-x86_64-linux \
    --option binary-caches 'https://cache.nixos.org/ https://yzhou216.cachix.org' \
    --option trusted-public-keys '
    cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
    yzhou216.cachix.org-1:O7nGwQnLGGBGE8zYFMipVzPVN956FcV57y9SqwEP+O8=
    ' \
    /nix/store/higickkqy6m3ivwm9dgkg1a0b9q6g0sd-qman-1.5.1

• aarch64-linux

  nix-store -r --add-root nixpkgs-pr-434438-aarch64-linux \
    --option binary-caches 'https://cache.nixos.org/ https://yzhou216.cachix.org' \
    --option trusted-public-keys '
    cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
    yzhou216.cachix.org-1:O7nGwQnLGGBGE8zYFMipVzPVN956FcV57y9SqwEP+O8=
    ' \
    /nix/store/0zp4r69myvx6my58m5igdk9r0j2g08fp-qman-1.5.1

• x86_64-darwin

  nix-store -r --add-root nixpkgs-pr-434438-x86_64-darwin \
    --option binary-caches 'https://cache.nixos.org/ https://yzhou216.cachix.org' \
    --option trusted-public-keys '
    cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
    yzhou216.cachix.org-1:O7nGwQnLGGBGE8zYFMipVzPVN956FcV57y9SqwEP+O8=
    ' \
    /nix/store/vripi293i6qcb90h81i72pmng81b81qn-qman-1.5.1

• aarch64-darwin

  nix-store -r --add-root nixpkgs-pr-434438-aarch64-darwin \
    --option binary-caches 'https://cache.nixos.org/ https://yzhou216.cachix.org' \
    --option trusted-public-keys '
    cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY=
    yzhou216.cachix.org-1:O7nGwQnLGGBGE8zYFMipVzPVN956FcV57y9SqwEP+O8=
    ' \
    /nix/store/x94qsyn88cygaszxpqj682jjqaj562kd-qman-1.5.1

---

x86_64-linux

✅ 1 package built:

• qman

---

aarch64-linux

✅ 1 package built:

• qman

---

x86_64-darwin (sandbox = true)

✅ 1 package built:

• qman

---

aarch64-darwin (sandbox = true)

✅ 1 package built:

• qman

[lilyball]

I don't believe it was ever necessary? AFAICT qman invokes all these tools via configured paths, defaulting to absolute paths. Certainly it seems to work fine for me without it.

[yzhou216]

I don't believe it was ever necessary? AFAICT qman invokes all these tools via configured paths, defaulting to absolute paths. Certainly it seems to work fine for me without it.

It fails in a pure shell:

$ nix-shell --pure -A qman
$ ./result/bin/qman
'apropos .' failed; did you run mandb/makewhatis?

Should we add it back or is there another way to do this?

[lilyball]

It fails in a pure shell

So does using nix-shell --pure -p man and running apropos .. Similarly, running env -i TERM="$TERM" PATH="$PATH" result/bin/qman still fails, which means it's not the PATH var that matters. I don't know enough about mandb to know what's going on, the only env vars listed in the man page for apropos are things I don't have set.

[lilyball]

Ah hah it's the $HOME env var. env -i TERM="$TERM" HOME="$HOME" PATH="$PATH" result/bin/qman works. Trying to do this in nix-shell --pure will just never work because it overrides PATH, which apropos uses to synthesize MANPATH if it's not otherwise set.

[lilyball]

This file is now semantically identical to what I proposed, so I'm happy with it.