- Introduce SourceT, which is simple variant of "correct `ListT`".
There are another variants possible (like in `streaming`),
but I'm not sure there's much real difference.
- Introduce `Codensity`. There's a flag if people don't want to depend
on `kan-extensions`.
- `StreamGenerator` and `ResultStream` are both `SourceT`.
`Stream` combinator in `servant-client` uses `Codensity` for CPS.
- Add servant-machines, servant-conduit, servant-pipes
- Add streaming cookbook: just code, no explanations.
- Add a script to run streaming 'benchmarks'
In case that a sub-server doesn't support the content-type specified
in the request invoke `delayedFail` instead of `delayedFailFatal` in
order to give the chance to other sub-servers to handle the request.
Changes Header, ReqBody and QueryParam to take a modifier list.
Resolves https://github.com/haskell-servant/servant/issues/856
ResponseHeader story turns to be somewhat ugly, but it can be made
elegant when https://github.com/haskell-servant/servant/issues/841 is
implemnted, then we can omit HList aka Header Heterogenous List
implementation.
- servant-server changes:
Writing server side intepretations is quite simple using
`unfoldRequestArgument`, which makes Header and QueryParam look quite
the same.
`ReqBody` cannot be easily made optional with current design (what that
would mean: No Content-Type Header?), so that dimensions isn't used
there.
- Add HasLink for all the rest ComprehensiveAPI combinators
- Add 'tricky' Header', QueryParam' endpoints to ComprehensiveAPI
- servant-docs: Quick'n'dirty implementation. Don't use modifiers information (yet).
For uniformity of Enter.
Previously, `ServerT Raw m ~ Application`. Seems reasonable, but has the
unfortunate consequence of making `Enter` useless for `Raw` routes.
With this change `Tagged m Application` is retagged by `Enter`.
This is a reasonably simple attempt at fixing #460.
By moving the accept check to a place before the body check,
we can make it recoverable (the body check is irreversible,
so everything done after the body check has to fail fatally).
The advantage is that we can now specify routes offering
different content types modularly. Failure to match one
is not fatal, and will result in subsequent routes being
tried.
The disadvantage is that we hereby bump the error priority
of the 406 status code. If a request contains a bad accept
header and a bad body, we now get 406 rather than 400. This
deviates from the HTTP decision diagram we try to follow,
but seems like an acceptable compromise for now.
