direwolf: refactor; direwolf-unstable: init at 1.7-unstable-2025-04-29#406106
direwolf: refactor; direwolf-unstable: init at 1.7-unstable-2025-04-29#406106SigmaSquadron merged 9 commits intoNixOS:masterfrom
Conversation
pkgs/by-name/di/direwolf/package.nix
Outdated
| --replace 'Terminal=false' 'Terminal=true' \ | ||
| --replace 'Exec=@APPLICATION_DESKTOP_EXEC@' 'Exec=direwolf' | ||
| substituteInPlace src/dwgpsd.c \ | ||
| --replace 'GPSD_API_MAJOR_VERSION > 11' 'GPSD_API_MAJOR_VERSION > 14' |
There was a problem hiding this comment.
This substituteInPlace doesn't match anything anymore. I assume this was to handle an incorrect version check, but I lack the context. Regardless, this should be safe to remove since it wasn't doing anything before anyway.
e7ebf22 to
76ae2d6
Compare
SigmaSquadron
left a comment
There was a problem hiding this comment.
versionCheckHook should just work because direwolf will always display its version when calling for --help (which versionCheckHook already does) or getting the help menu after passing an invalid flag.
You can also set versionCheckProgramArg to version, which is a command in direwolf, but that shouldn't be necessary either.
| }).overrideAttrs | ||
| (finalAttrs: { | ||
| # Next version is 1.8 | ||
| version = "1.8-unstable-2025-04-29"; |
There was a problem hiding this comment.
| version = "1.8-unstable-2025-04-29"; | |
| version = "1.7-unstable-2025-04-29"; |
Please review the package versioning guidelines:
- If we're tracking a commit from a repository without a version assigned, then the version attribute should be the latest upstream version preceding that commit, followed by
-unstable-and the date of the fetched commit. The date must be in ISO 8601 "YYYY-MM-DD" format.
That's to say, if 1.8 isn't tagged, it can't be added here.
There was a problem hiding this comment.
Huh, I missed that. I'll fix gpredict-unstable's version in a seperate PR then.
There was a problem hiding this comment.
I'm not sure we should be adding unstable package versions to nixpkgs at all - this seems like a NUR thing.
There was a problem hiding this comment.
Both have very slow release cycles and have some significant features that I would like to use that aren't present in the latest stable version. This shouldn't require much maintenance, considering updates are going to be, for the most part, completely automated. If there is a policy or other concrete reason against including unstable variants of packages, then fair enough, but both are, in my opinion, value added.
I do get what you mean about it maybe being a better fit for NUR. In fact, I do have a custom direwolf derivation in my own nur-packages. I feel that that shouldn't preclude it from being in nixpkgs though. Nixpkgs has a higher standard for quality than NUR does, and I believe this meets that higher standard.
There is no |
|
Seems like setting |
|
Actually it looks like the |
76ae2d6 to
cad686d
Compare
cad686d to
0b5cf01
Compare
| [ | ||
| "deviceid.c" | ||
| "tocalls.yaml" | ||
| "--replace-fail /usr/lib/udev/rules.d/ ${placeholder "out"}/lib/udev/rules.d/ --replace-fail /etc/udev/rules.d/" |
There was a problem hiding this comment.
This is pretty ugly but I don't see an easier way
|
Why is the nixfmt job angry at some random python package?! |
|
someone might have merged a broken PR :p Edit: yeah it was #406492 |
|
Spurious newline, sorry. Fix is in #406498 |
0b5cf01 to
30c304f
Compare
|
Rebased |
|
Haven't looked at this in depth, only added myself as maintainer because there was no one else at the time. But I was the one who added gpsd support initially, and someone else disabled it by default due to frequent breakages in the gpsd API for what that's worth. It seems like a fairly marginal feature not worth breaking the package on a regular basis IMO, unless someone is really on top of maintaining it. |
SigmaSquadron
left a comment
There was a problem hiding this comment.
Diff wise it seems fine to me! Love to see the split commits.
I set
nixpkgs/pkgs/by-name/gp/gpsd/package.nix Line 20 in db8c5bd |
30c304f to
1bf596f
Compare
|
Rebased and removed the commit enabling |
sarcasticadmin
left a comment
There was a problem hiding this comment.
Both version build for me
Thanks for keeping the original default @Pandapip1 ![]()
|
Oh, whoops. I forgot about this PR. @Pandapip1: I've already approved this; if this is still necessary, just rebase and I can merge it. |
b71e597 to
bc3370f
Compare
|
Actually, it needs formatting. |
|
Fixed. |
|
Squash the formatting commit. |
0bfb151 to
f577938
Compare
|
I hope I resolved those merge conflicts correctly. |
|
Please format the derivation. The CI is still failing. |
|
Sorry, every time I have to make a change like that I have to manually resolve at least three merge conflicts. |
f577938 to
25e32a4
Compare
A workaround was foundversionCheckHookwas not added due to the lack of a--versionsubcommand. See wb2osz/direwolf#570Things done
nix.conf? (See Nix manual)sandbox = relaxedsandbox = truenix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)Add a 👍 reaction to pull requests you find important.