* This fails two of the tests. We should probably add a catch all
at the end that captures anything not matches by the Symbols for
the preceding branches `arms` and `legs`.
* When using `CaptureAll` from the root "/" and empty path becomes
an empty list. when using it from a branch though an empty path,
ie. a trailing separator, becomes `mempty` for the type. Where is
this happening? I think this is definitely happening before
`parseUrlpieces` gets run, so it happens in the machinery for
Delayed and DelayedIO. I am still looking for the code that
produces the `txts` argument for the `captureD` lambda created in
the `HasServer` instance for `CaptureAll`.
Close#1513.
GHC 9.2 needs explicit kind signature here, I don't really understand
why.
This kind signature is correct and not too restritive, because `HasLink`
is technically defined `class HasLink endpoint` which means that it is
infered as `k -> Constraint`. In the instance signature, we have
`HasLink ((arr :: a -> b) :> sub)`, so here the `k` is the same kind as
the one of `:>` which is not polykinded.
As the head method isn't allowed to contain any response body, no
general Head Verb is added. (This may easily lead to wrong usages...)
(https://httpwg.org/specs/rfc7231.html#HEAD)
* bumped cabal-version field
Cabal supports two types of licenses, native and SPDX, which can be seen here hackage.haskell.org/package/Cabal-3.6.2.0/docs/Distribution-Types-PackageDescription.html#v:licenseRaw
Several packages use BSD-3-Clause as a license, in conjonction with cabal-version: >=1.10 which cabal parses as Right (UnknownLicense "BSD-3").
If I change teh cabal-version to cabal-version: 2.2 , cabal correctly identifdies the license License (ELicense (ELicenseId BSD_3_Clause)).
* changed license from cabal to spdx format
aka BSD3 -> BSD-3-Clause: next cabal may deprecate the old format
Move `HasServer (NamedRoutes routes)` instance
The instance has been moved to `Servant.Server.Internal`, as the
instances for other combinators. It is necessary so that the instance
can be re-exported from `Servant.Server` without circular imports.
Otherwise, users have to import `Servant.Server.Generic` manually ;
forgetting to do so will produce confusing error messages about the
missing instance.
Move `HasClient (NamedRoutes routes)` instance
Moved so that the instance is made available when importing
`Servant.Client`, avoiding possibly confusing errors when
`Servant.Client.Generic` isn't imported.
QuantifiedConstraints isn't available for GHC 8.4 (where our GHCJS
version is still stuck).
We may need to take a drastic decision for GHCJS at some point.
We define `ServerT (NamedRoutes api) m` as `api (AsServerT m)`, so that
the server of an record-defined API is a record of handlers.
The implementation piggy backs on the instance for “vanilla” servant
types with `(:<|>)`, using the `GServantProduct` for converting backd
and forth between the record / vanilla servers.
The main difficulty is that GHC needs to know that this operation is
legit, which can be expressed as the fact that:
```
GToServant (Rep (ServerT (NamedRoutes api))) m ~
ServerT (GToServant (Rep (api AsApi))) m
```
plus a few additional constraints.
This is easy enough for `route`, as we know that `m ~ Handler`. But in
the case of `hoistServerWithContext`, the two involved monads are
unknown ; in other words, this constraint needs to hold `forall m.`
Switching `-XQuantifiedConstraints` on is not sufficient, as our
constraints involve type families (`Rep` and `ServerT`). Our trick is to
use an intermediary typeclass, `GServer`, as a provider of evidence (in
the form of a `Dict`) that our constraints are indeed satisfied for a
particular monad.
The only instance of `GServer` is defined along with it, so it is
practically invisible to users.
Allows users to directly embed APIs defined as records of routes into
vanilla Servant API types.
E.g.:
```haskell
data MyRoutes mode = MyRoutes
{ version :: mode :- Get '[JSON] Int
, …
}
type API = "prefix" :> NamedRoutes MyRoutes :<|> …
```
APIs can thus be recursively defined directly with Generic record types.
ServerError field errBody uses ByteString, whose IsString instance kills
Unicode, thus turning example into garbage. Changed it to simple ASCII
string, since Unicode art did not exactly correspond to 404 error
anyway.
Fixes#1371
