val jwtStatelessAuthenticator =
JWTAuthenticator.pstateless.inHeader[F, User, HMACSHA256](
signingKey = signingKey,
settings = TSecJWTSettings(JWT_AUTH_HEADER, FiniteDuration(24L * 3600L, TimeUnit.SECONDS), None)
)
val authUser: Kleisli[OptionT[F, ?], Request[F], User] = Kleisli({ req =>
if (userAuthentication) {
jwtStatelessAuthenticator
.extractAndValidate(req)
.map(_.identity)
} else {
OptionT.pure[F](User.dummy)
}
})
private[auth] val defaultNotAuthenticated: AuthedRequest[F, String] => OptionT[F, Response[F]] =
req => OptionT.liftF(Forbidden(req.authInfo))
val onAuthFailure: AuthedRoutes[String, F] = Kleisli(defaultNotAuthenticated)
val authenticated: AuthMiddleware[F, User] =
AuthMiddleware.withFallThrough(authUser)
F[_] instead of IO. Is RhoRoutes suitable for that? I guess it is, although all the example code / documentation uses IO. The reason I'm asking for is I have a route that returns different status codes based on some intermediate results and it works fine with IO but I can't get it to be of the required F[Result[...]] type with F[_]:ConcurrentEffect, instead it just defaults to Any As far as Intellij's shw implicits feature goes I can't see any difference. My ConcurrentEffect is being picked up instead of ioEffect which seems to be fine.
Hello there. Is it possible to get this feature backported to the 0.19.x releases series? We are stuck using the 0.20.x Http4s series right now, and we're trying to build some functionality that depends on the content-type headers:
Happy to open a PR against a branch if one is opened for me.
@zarthross is this something you can help me out with? We're using Rho internally but this is kind of a roadblock for us. We want to have a route support multiple content-types (e.g. json and proto), and we want to use the content-type header to control that. The Swagger document will unfortunately not work properly for this use case without that backport.
We are using 0.19.x because our team is still supporting Scala 2.11 at this time (and that cannot be changed until other vendor changes are completed).
Hello, I updated ther dependencies in my project,
cats-core = 2.1.1
cats-effect = 2.1.2
rho-swagger = 0.20.0
http4s-blaze-server = 0.21.0
now, I have a problem in start the project, the error is
Could not find an instance of Applicative for [x]F[x]
[error] server <- BlazeServerBuilder[F]
[error] ^Any suggestions to solve this error?, The code is this
import cats.effect.{ConcurrentEffect, ContextShift, ExitCode, IO, IOApp, Resource, Timer}
import co.yyyy.xxxxAPI
import org.http4s.server.Server
import org.http4s.server.blaze.BlazeServerBuilder
object Main extends IOApp {
def run(args: List[String]): IO[ExitCode] =
AppServer.create.use(_ => IO.never).as(ExitCode.Success)
}
object AppServerSwagger {
def create[F[+_]: ConcurrentEffect: ContextShift: Timer]: Resource[F, Server[F]] =
for {
server <- BlazeServerBuilder[F]
.withHttpApp(xxxxAPI[F].httpApi)
.bindLocal(8085)
.resource
} yield server
}Any error message
[error] found : cats.effect.Resource[[x]F[x],Nothing]
[error] required: cats.effect.Resource[F,org.http4s.server.Server[F]]
[error] config <- Resource.liftF(parser.decodePathF[F, DefaultConfig]("app"))
[error] ^
pathVars not supported like query params in terms of Enums? See above. Both use a StringParser based on an Enumeratum enum, but the path param comes through odd
implicit def producerTypeStringParser[F[_]]: StringParser[F, ProducerType] =
new StringParser[F, ProducerType] {
override def parse(
s: String
)(implicit F: Monad[F]
): ResultResponse[F, ProducerType] =
try SuccessResponse(ProducerType.withName(s))
catch {
case _: NoSuchElementException =>
FailureResponse.pure[F] {
BadRequest.pure(s"Invalid ProducerType format: '$s'")
}
}
override val typeTag: Option[TypeTag[ProducerType]] = Some(
implicitly[TypeTag[ProducerType]]
)
}
val myIdsList = paramD[List[Long]]("myIds", "Some ids") which works as expected: it parses ?myIds=1&myIds=2 to a List[Long]. However, how can I use my own StringParser[IO, List[Long]] instead of QueryParsers.multipleParse? I'd like to manually parse ?myIds=1,2 to List[Long].
The above came up during an upgrade from v0.19.0 to v0.20.0 where I noticed that suddenly, our custom StringParser[IO, List[Long]] wasn't picked up and instead, QueryParsers.multipleParse was being used.
What I see is that v0.19.0's multiParse() is defined for Seq[_] while v0.20.0's multiParse() is defined for generic B[_].
I also see that QueryParsers.standardCollector and QueryParsers.multiParse are defined in the same scope. Would it help here if standardCollector had higher priority than multiParse()?
multiParse by defining the following: implicit def http4sMultipleParseOverride[A](
implicit F: Monad[IO],
p: StringParser[IO, List[A]],
cbf: Factory[A, List[A]]
): QueryParser[IO, List[A]] =
standardCollector[List[A]](F, p)
As an alternative to
'foo we've introduced pv"foo". pv stands for pathVar.You can see an example here https://github.com/http4s/rho/blob/master/examples/src/main/scala/com/http4s/rho/swagger/demo/MyRoutes.scala#L45
logger ?
Hi @Yotamho
You cloud copy the source of RhoRoutes, and have your own version with the logger field removed. That's what I would do.
Among other options you can also create a block { } inside your class extending RhoRoutes. And while you cannot create a 'logger' field in the class, you can create a val logger in the block, which will shadow the logger field inherited from RhoRoutes. You would then have to put all your code using the logger inside that block. That would work as long as you don't plan having any public members in that class (you can't put public members in the block). It's a bit hacky :P
- in: "body"
name: "body"
description: "List«Long»"
required: true
schema:
type: "array"
items:
title: "Long"
$ref: "#/definitions/Long"