summaryrefslogtreecommitdiff
path: root/src/ui/mod.rs
Commit message (Collapse)AuthorAge
* Handlers are _named operations_, which can be exposed via routes.Owen Jacobson2025-06-18
| | | | | | Each domain module that exposes handlers does so through a `handlers` child module, ideally as a top-level symbol that can be plugged directly into Axum's `MethodRouter`. Modules could make exceptions to this - kill the doctrinaire inside yourself, after all - but none of the API modules that actually exist need such exceptions, and consistency is useful. The related details of request types, URL types, response types, errors, &c &c are then organized into modules under `handlers`, along with their respective tests.
* Reorganize and consolidate HTTP routes.Owen Jacobson2025-06-18
| | | | | | | | HTTP routes are now defined in a single, unified module, pulling them out of the topical modules they were formerly part of. This is intended to improve the navigability of the codebase. Previously, finding the handler corresponding to a specific endpoint required prior familiarity, though in practice you could usually guess from topic area. Now, all routes are defined in `crate::routes`. Other than changing visibility, I've avoided making changes to the handlers at the ends of those routes.
* Unify `setup_required` middlewares.Owen Jacobson2025-06-17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The two middlewares were identical but for the specific `IntoResponse` impl used to generate the response when setup has not been completed. However, unifying them while still using `from_fn_with_state` lead to this horrorshow: .route_layer(middleware::from_fn_with_state( app.clone(), |state, req, next| { setup::middeware::setup_required(UNAVAILABLE, state, req, next) } )) It's a lot to read, and it surfaces the entire signature of a state-driven middleware `fn` into the call site solely to close over one argument (`UNAVAILABLE`). Rather than doing that, I've converted this middleware into a full blown Tower middleware, following <https://docs.rs/axum/latest/axum/middleware/index.html#towerservice-and-pinboxdyn-future>. I considered taking this further and implementing a custom future to remove the allocation for `Box::pin`, but honestly, that allocation isn't hurting anyone and this code already got long enough in the translation. The new API looks like: .route_layer(setup::Required::or_unavailable(app.clone())) Or like: .route_layer(setup::Required::with_fallback(app.clone(), RESPONSE)) One thing I would have liked to have avoided is the additional `app.clone()` argument, but there isn't a way to extract the _state_ from a request inside of an Axum middleware. It has to be passed in externally - that's what `from_fn_with_state` is doing under the hood, as well. Using `State` as an extractor doesn't work; the `State` extractor is special in a _bunch_ of ways, and this is one of them. Other extractors would work. Realistically, I'd probably want to explore interfaces like .route_layer(setup::Required(app).or_unavailable()) or .route_layer(app.setup().required().or_unavailable())
* Set `charset` params on returned content types.Owen Jacobson2024-10-22
| | | | This is a somewhat indirect change; it removes `mime_guess` in favour of some very, uh, "bespoke" mime detection logic that hardcodes mime types for the small repertoire of file extensions actually present in the UI. `mime_guess` doesn't provide a way to set params as it exports its own `Mime` struct, which doesn't provide `with_params()`.
* Organizational pass on endpoints and routes.Owen Jacobson2024-10-16