Thanks for your contribution! I only realized now that the comments below have reached double digits, but I hope you don't mind that these are what brings your contributions more in line with what Nixpkgs expects.

I haven't tried it out yet, and I don't fullyunderstand how the build process works, so hopefully someone with more energy or familiarity can provide further insights.

In addition to code suggestions, the commit message for the second commit should be dictdDBs.gcide: init at 0.53. Also please remember to squash any fixups at the end, so the history remains clean.

Thanks for the rapid && actionable feedback! Oh and mind it I do not as main aim is to contribute, while also not making a mess ;-)

Build process is totally stolen from the AUR dict-gcide, including the sed-scripts, because I use Arch BTW™ but am making a transition to NixOS. Two things, so far, worthy of note;

• Git repo from AUR does not include Git tags or similar, so updates here are likely to lag
• It might have been smarter, and in the future/present smarter, to instead clone directly from AUR and run PKGBUILD script, then move stuff after

As for squashing and updating commit message, I'll be sure to do that once things are in an acceptable state, thanks for the correction(s)!

Update 2025-07-25 04:06 UTC: I've mostly got change requests sorted... mostly... And current commit changes are targeted to hopefully make reviewing through time a bit easier. The version attribute and libmaa dependency are where I'm expecting are still wrong.

I want to clarify before any misunderstanding that by squash, I mean eventually getting down to three commits: Adding yourself to maintainers list, extracting libmaa, dictdDBS.gcide: init

I want to clarify before any misunderstanding that by squash, I mean eventually getting down to three commits: Adding yourself to maintainers list, extracting libmaa, dictdDBS.gcide: init

Ah, okay that seems reasonable, and I'll get on re-ordering changes and squashing thangs

Update 2025-07-25 08:24 UTC: I've re-ordered and squashed and updated commit messages 🎉

Currently I'm doing a fsck on the device hosting my NixOS install, had a hard-shutdown, so testing on my end if libmaa really is finally available via function arguments... and that I didn't break dictd... may be a few hours away x-)

Thanks dramforever for sticking with me, and ensuring I don't make a total mess of things!

Latest force-push (commit hash 3b6c7f24a0b6) should include almost all change requests, and local build/tests seem to be okay 🎉

[root@nixos:/etc/nixos]# nixos-rebuild switch --impure --flake . && dict hello

warning: Git tree '/etc/nixos' is dirty
building the system configuration...
warning: Git tree '/etc/nixos' is dirty
#> ...
Done. The new configuration is /nix/store/377vr3j69x4kbhmkgd777pabyh7l12kv-nixos-system-nixos-25.11.20250708.9807714
1 definition found

From The Collaborative International Dictionary of English v.0.53 [dict-gcide]:

  Hello \Hel*lo"\, interj. & n.
     An exclamation used as a greeting, to call attention, as an
     exclamation of surprise, or to encourage one. This variant of
     {Halloo} and {Holloo} has become the dominant form. In the
     United States, it is the most common greeting used in
     answering a telephone.
     [1913 Webster +PJC]

The sed scripts have unclear license. Shipping it in Nixpkgs would amount to relicensing it to MIT. We should ask the AUR author about it.

Smart! I figured the "Package Details" → "Licenses" value was the project's license, but also get how even though it's a copy-left license it could technically apply to the package being distributed only.

Consider it asked :-)
And I got notifications turned-on, so if/when the team answers I'll be sure to

In general, a fooPhase needs runPhase preFoo at the start and runPhase postFoo at the end. This is a limitation in the current phase framework of Nixpkgs, but it is what it is.

🤔 I'm not sure if this is something I need to take action on, because I moved stuff to the prePatch phase as requested, but if it be more of a keep it in mind for any future packaging then I certainly will. And am grateful regardless!

No ... here please. Again, check how it's done on similar packages.

Ooof, okay that's a bit embarrassing 🤦
Here I thought I read the ellipsis was needed to prevent errors of too many args, but nope! And re-looking at neighboring packages I should've been smorter

Latest force push will have that removed as well as missing semicolons at the checkPhase added. Please do let me know if there's anything else I gotta tackle while with await an answer from AUR maintainers!

In general, a fooPhase needs runPhase preFoo at the start and runPhase postFoo at the end. This is a limitation in the current phase framework of Nixpkgs, but it is what it is.

🤔 I'm not sure if this is something I need to take action on, because I moved stuff to the prePatch phase as requested, but if it be more of a keep it in mind for any future packaging then I certainly will. And am grateful regardless!

I should have spelled it out, but in any case I simply meant installPhase and buildPhase

In general, a fooPhase needs runPhase preFoo at the start and runPhase postFoo at the end. This is a limitation in the current phase framework of Nixpkgs, but it is what it is.

🤔 I'm not sure if this is something I need to take action on, because I moved stuff to the prePatch phase as requested, but if it be more of a keep it in mind for any future packaging then I certainly will. And am grateful regardless!

I should have spelled it out, but in any case I simply meant installPhase and buildPhase

Heh, thanks! I can be kinda dense, sometimes, and will for sure spend some quality time with the docs before submitting any new PRs

Edit 2025-07-29 05:36 UTC -- Uh, so skimming some other packages I'm seeing runHook preInstall within installPhase and, again, feel like I gotta ask if ya be asking for changes? Like should I do runPhase checkPhase within the installPhase instead of leveraging the doCheck = true attribute?

   '';

-  doCheck = true;
   checkPhase = ''
@...@
   '';

   installPhase = ''
+   runPhase checkPhase
    install -Dm 0644 -t "$out/share/dictd/" ./gcide.{dict.dz,index};

Thought it wise to pop-in, no updates from the AUR package maintainers about license of .sed scripts, but I did to a bit of exploring of commit history via git whatchanged *.sed

And found there be one author that is involved with those files, who was last active just under one year ago for that repository. Plus a little cyber-stalking, AKA searching the commit email address on GitHub, shows @respiranto might be who we need to ask licensing clarification of.

Note I base this on checking the commit history of;

https://github.com/lawandorga/lawandorga-mail-server/commit/b900def64fe393f638f3195fd0e5e66c96c2d871

... and spotting the same name/email combo as what's in the AUR

Though I am not sure about the etiquette of reaching out directly via email just yet, maybe a public ping here is better 🤷

I have added licensing information on the sedfiles and such in the AUR package.

@respiranto you're the best! I noticed in the diff some things were removed, other than comments, and a few executable lines were added too. But rebuild against version 0.53 seems to be functional, and run-time is happy, so I'm gonna trust ya ;-)

@dramforever I tried bumping the version locally to 0.54 but that led to fetchurl time-outs so, for now, I'm gonna keep the proposed version at 0.53 for folks to enjoy and report back on. Please do let me know, when ya got a chance, if anything else need be done by me to get this PR into an acceptable state.

@S0AndS0:

• You may have used an older version of the AUR package; my commits from today definitely only added licensing. The previous commit is from 2025-07-14.
• The current state of the sedfiles is specifically adapted to version 0.54; it may work for 0.53, but I actually removed many no-longer-needed fixes on update for version 0.54.

Ah thanks for clarifying where other changes be from, and I'll try bumping the version here tomorrow to 0.54 as it my have been a rate-limiting thing that caused time-outs 🤷

Update 2025-08-14 20:36 UTC

I was being dumb about the time-outs 🤦 version and hash be updated in the latest force-push and I'll get to updating the title for this PR in a moment