Configures the compute thread pool used to process some async computations.
Configures the compute thread pool used to process some async computations.
This defaults to cats.effect.Async[F].executionContext. In
almost all cases, it is desirable to use the default.
The Blaze server has a single-threaded event loop receiver used
for picking up tcp connections which is completely separate to
this pool. Following picking up a tcp connection, Blaze shifts
to a compute pool to process requests. The request processing
logic specified by the HttpApp is executed on the
cats.effect.Async[F].executionContext. Some of the other async
computations involved in request processing are executed on this
pool.
Sets the handler for errors thrown invoking the service. Is not
guaranteed to be invoked on errors on the server backend, such as
parsing a request or handling a context timeout.
Sets the handler for errors thrown invoking the service. Is not
guaranteed to be invoked on errors on the server backend, such as
parsing a request or handling a context timeout.
Returns the backend as a single-element stream. The stream
does not emit until the backend is ready to process requests.
The backend is shut down when the stream is finalized.
Returns the backend as a single-element stream. The stream
does not emit until the backend is ready to process requests.
The backend is shut down when the stream is finalized.
Returns an effect that allocates a backend and an F[Unit] to
release it. The returned F waits until the backend is ready
to process requests. The second element of the tuple shuts
down the backend when run.
Returns an effect that allocates a backend and an F[Unit] to
release it. The returned F waits until the backend is ready
to process requests. The second element of the tuple shuts
down the backend when run.
Unlike resource and stream, there is no automatic
release of the backend. This function is intended for REPL
sessions, tests, and other situations where composing a
cats.effect.Resource or fs2.Stream is not tenable.
resource or stream is recommended wherever possible.