-
Notifications
You must be signed in to change notification settings - Fork 13
decisions #2
Comments
Great work so far! I added three more decisions to be done (REST+API blueprint+Testing framework). Let us discuss today who will work on that. |
After further research and from some little experience I have it seems Jest is a common standard for testing for React apps and it is also a recommended tool by React. I don't think a decision here is necessary, what do you think? |
I agree, no decision is needed but there should be a short documentation about the testing strategy. |
Sounds good to me!
Can I get more context on API Blueprint please, do you mean this? |
Exactly, then we can use like https://apiblueprint.org/tools.html e.g. for testing in the CI. Doesn't seem to have Rust generators, though. Maybe it is over-engineering to use something like this. |
I see, it is not needed then for now I suppose? Do you have a suggestion on how in-depth the testing strategy documentation should be/what it should include? |
Yes, if you document the API properly as code comments it is fine by me, too.
It doesn't need to be very in-depth, it should mostly cover:
At the moment very rough descriptions are enough, as we are not finished with the use cases yet. |
Someone told me about tauri-apps/tauri#1514 (Tauri itself might also be interested in the long-term), which discusses type-safety between Rust and Typescript. There seem to be quite a few options:
Can you please write up a decision, even if we decide to not do anything of it. |
The main architecture pattern for the frontend is also to be decided, e.g.:
The decision will also influence how undo is done, e.g. by using https://www.npmjs.com/package/redux-undo-redo-middleware? |
If we are using JSON schema, another option would be: https://imfeld.dev/writing/generating_typescript_types_from_rust |
I played around with a bit and it seems like typeshare is the one that works the best seamlessly, it is also easy to configure so I'm leaning towards using it |
LGTM, too. |
@buenaflor if you want to look into graphql anyway, it would be nice to have a short write-up what you found in a decision. (If you don't look, don't write 😉). Some libs seem to be a bit immature but there already seems to be everything needed:
It is not totally unrealistic that for lookups in the map REST is not ideal (slow). |
We got very positive feedback about our decisions: "React + Vite + TypeScript klingt gut, das ist aktuell der populärste & stabilste Stack für Frontend Projekte" For State Management, however, we should look into:
I updated the top post. |
There is already a small state-management solution in use. It's called zustand. If we switch to a different library we should get rid of this dependency. |
@absurd-turtle exactly, that is what is to be decided. See #30 for prior work, I think you can take over the work on this PR. |
@ElektraInitiative/permaplant
If you want one of them assigned, please write here. |
@markus2330 about the rss generator: What should I write there and where? However there isn't really a decision involved. It is more 'How to implement a RSS generator' so it doesn't really fit in decisions. |
The idea was to have some Markdown file (or something else easy to write manually) that is used for both:
Ideally we would use something that already fully implements this, which doesn't seem to exist in Rust. @absurd-turtle can you maybe share your idea? Would the backend be involved? Is there some ready-made solution. |
I think the best solution is to generate the RSS feed in the backend and serve the XML on a specific endpoint. To store the data we could just have files that are served directly or put it in the database. |
@absurd-turtle As the news feed on the web site should now be a NC chat, it would be great if we can generate the RSS feed from the chat's content. Is this better implemented in the frontend or in the backend? |
We could try to use
But it is a bit much experimental technology involved but maybe okay for a public announcement channel. Having our announcement channel in many formats would not be a bad thing. |
soon
to have a look at:
if this is easily possible, we will:
probably decided
done
The text was updated successfully, but these errors were encountered: