draupnir: init at 2.0.0-beta.6 #342893
draupnir: init at 2.0.0-beta.6 #342893
Conversation
|
|
||
| # TODO: move to preNpmRebuild | ||
| preBuild = '' | ||
| rm -r node_modules/better-sqlite3/{build,deps/common.gypi,deps/sqlite3.gyp,deps/sqlite3} |
There was a problem hiding this comment.
# Remove things we know we don't need that just increase closure size.
There was a problem hiding this comment.
Isn't that still needed for sqlite?
There was a problem hiding this comment.
I believe the removed files are all only used by the better-sqlite3 package to build it's own vendored sqlite. better-sqlite3-dynamic-link.patch replaces that with just linking against the nixpkgs sqlite.
| nativeBuildInputs = [ pkg-config ]; | ||
| buildInputs = [ sqlite ]; | ||
|
|
||
| # TODO: move to preNpmRebuild |
There was a problem hiding this comment.
Unsure yet if I'd like to do this in this PR (would decrease build time by a bit), but we'll see after it gets tested.
| uses across an entire homeserver. | ||
| ''; | ||
| license = licenses.afl3; | ||
| maintainers = with maintainers; [ winter ]; |
There was a problem hiding this comment.
@TheArcaneBrony Do let me know if you'd like to be added as a maintainer.
ofborg
bot
added
8.has: package (new)
There was a problem hiding this comment.
IIRC vendored package-lock.json was why the Draupnir package was rewritten from scratch
Using this package directly with the mjolnir module isn't quite drop-in because the mjolnir module doesn't have a
The mjolnir default config is compatible with draupnir, so we ended up just doing systemd.services.mjolnir.serviceConfig.ExecStart = lib.mkForce ''${pkgs.draupnir}/bin/draupnir --draupnir-config ./config/default.yaml'';With this hack, we're running the package on @scd31's infrastructure. It appears to be working so far. Will see if anything breaks and come back with more details. |
|
Thank you! |
|
I may try to write a proper NixOS test for this some time this weekend for the hell of it. |
|
nvm this looks suspiciously not possible. but maybe it can still be done by other means?? |
|
we could but i think it'd be way too fragile dependency tree resolving wise (e.g. yarn may produce a diff. tree that npm would need to call out to the internet to fix etc.) |
ideally we could convince npm to just accept that it got a different solution than it would have dreamt up itself, but yeah it is a little sketchy. relatable problems with dealing with flakes so i get it. just wish there was a better way somehow. |
|
dynamic drvs will save us all :( |
|
Worst case you can use the draupnir module from my package and specify the |
Description of changes
This supersedes the package portion of #274052, as discussed in Matrix, by packaging Draupnir in a way that is:
mkYarnPackage, thus requires slightly less work, like no manual wrapping being required).However, this has the caveat that it requires a vendored
package-lock.json. Sorry!This needs testers! I'm not sure if I can just drop in the Mjölnir module and test it or not, though I guess someone can just test it with the module from #274052.
Thanks to @TheArcaneBrony for creating the original PRs and the module, which will continue to be worked on in that PR.
Things 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.