emanote: init at 0.8.0.0#199442
Conversation
|
Fun fact: For testing this I started emanote in the nixpkgs root, because that‘s where I had the build result. And that actually proved quite useful and I could read the nixpkgs docs via the emanote webserver. Anyway, @srid I am a bet weary about the tailwindcss overrides. I have just copied them, but do you maybe have a bit of documentation for why those are needed. I looked for that in the tailwind README on hackage. I don‘t like the idea, that this is bespoke knowledge only present in your flake and nixpkgs. Besides that I haven‘t done anything about M1, yet. |
srid
left a comment
There was a problem hiding this comment.
Nice!
I assume if I use emanote as a Haskell dependency in .cabal, all of this configuration (tailwind, stork dependency) apply there as well?
There was a problem hiding this comment.
I think we can remove "@tailwindcss/language-server". But the other plugins are generally used. Pretty much anything you can use in the Emanote HTML templates as a user can be added here.
There was a problem hiding this comment.
Anyway, @srid I am a bet weary about the tailwindcss overrides. I have just copied them, but do you maybe have a bit of documentation for why those are needed. I looked for that in the tailwind README on hackage. I don‘t like the idea, that this is bespoke knowledge only present in your flake and nixpkgs.
I should probably expand the README of https://github.com/srid/tailwind-haskell .... but since tailwind Haskell package is responsible for mainly compiling the classes to produce a final .css, I would say depending on all plugins that are useful in that compilation step is what is relevant. That means (AFAIU) all the plugins here except langauge-server.
I believe it should just work. EDIT: For intel mac (which I do not own to test), we want to remove the "broken" flag for |
This comment was marked as outdated.
This comment was marked as outdated.
Okay, it worked on M1 on the second run. Looks like the |
|
I think the stork problem on M1 is kinda orthogonal to this PR. I cannot test or verify it. I invite someone with darwin stakes to mark it unbroken. |
|
Tested emanote, seems to work. The closure size is still enormous, but I guess that's generally the case with Haskell executables? |
Yea, do we know what the standard practice on nixpkgs is to reduce the size of final Haskell packages (in cf. srid/emanote#295 |
|
Ah, yeah, we should probably add the remove-references-to. |
|
FWIW, this is what I did for neuron: |
|
Honestly, the closure size for nix builds of emanote have been blocking me from deploying it for awhile, so it would be amazing if |
|
and.. thanks for your work on this |
|
@edrex I would very much welcome your help! Would especially nice to get this in before branch-off on the 21st. |
|
Ok, I'll follow up with a PR when i have a draft. |
|
I dropped the ball on this and missed the branch-off window - been working more with a hammer than a keyboard these past weeks. Still top of queue for nix stuff. |
|
No worries. We can also always do a backport for this. So no meet to worry about branch off. |
|
Emanote 0.8.2 will also be soon released (with these changes) marking the official public initial release. So there is another opportunity, @edrex EDIT: Maybe I should call it 1.0. |
Description of changes
Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)nixos/doc/manual/md-to-db.shto update generated release notes