import com.evolutiongaming.catshelper.ClockHelper._
val clock = Clock.const[Id](nanos = 1000, millis = 2)
clock.millis // 2
clock.nanos // 1000
clock.micros // 1
clock.instant // Instant.ofEpochMilli(2)
Provides a way to measure duration of a computation in a pure way.
Example:
import com.evolutiongaming.catshelper.MeasureDuration
for {
duration <- MeasureDuration[IO].start
_ <- doSomething
duration <- duration
} yield duration
Syntax extensions are also available, allowing to measure duration of a computation and execute an effect with it:
import com.evolutiongaming.catshelper.syntax.measureDuration._
for {
int1 <- IO.pure(1).measured(elapsed => IO.println(s"elapsed: $elapsed"))
int2 <- IO.pure(1).measuredCase(
successF = elapsed => IO.println(s"Succeeded: $elapsed"),
failureF = elapsed => IO.println(s"Failed: $elapsed")
)
} yield int1 + int2
Like Ref
but allows A => F[A]
rather than A => A
Ensures that updates are run serially
import com.evolutiongaming.catshelper.SerialRef
for {
ref <- SerialRef.of[IO, Int](0)
_ <- ref.update(a => (a + 1).pure[IO])
} yield {}
Functional alternative to lazy
keyword in Scala
trait LazyVal[F[_], A] {
def get: F[A]
def getLoaded: F[Option[A]]
}
trait ToFuture[F[_]] {
def apply[A](fa: F[A]): Future[A]
}
trait FromFuture[F[_]] {
def apply[A](future: => Future[A]): F[A]
}
trait ToTry[F[_]] {
def apply[A](fa: F[A]): Try[A]
}
trait FromTry[F[_]] {
def apply[A](fa: Try[A]): F[A]
}
trait Log[F[_]] {
def debug(msg: => String): F[Unit]
def info(msg: => String): F[Unit]
def warn(msg: => String): F[Unit]
def warn(msg: => String, cause: Throwable): F[Unit]
def error(msg: => String): F[Unit]
def error(msg: => String, cause: Throwable): F[Unit]
}
trait Runtime[F[_]] {
def availableCores: F[Int]
def freeMemory: F[Long]
def totalMemory: F[Long]
def maxMemory: F[Long]
def gc: F[Unit]
}
trait ThreadLocalRef[F[_], A] {
def get: F[A]
def set(a: A): F[Unit]
def update(f: A => A): F[Unit]
def modify[B](f: A => (A, B)): F[B]
}
This is useful to ensure release
called at most once, in cases when "unsafe" api like Resource.allocated
being used
val resource: Resource[F, A] = ???
resource.fenced
A mutable reference to A
value with read-write lock semantics.
Manages a given Resource[F, A]
providing access to it only when a feature-toggle is on.
val serviceResource: Resource[F, AService] = ???
val flag: F[Boolean] = ???
val ftService: Resource[F, Resource[F, Option[AService]]] = FeatureToggled
.polling(
serviceResource,
flag,
pollInterval = 10.seconds,
gracePeriod = 30.seconds,
)
ftService.use { access =>
access.use {
case Some(service) => service.doStuff(…)
case None => F.unit
}
}
The logback module lives in a separate `cats-helper-logback' module to avoid dependency on a logback in case the user chooses a different logging backend. This is important to avoid the problem of multiple bindings when mapping the logging framework with SLF4J.
Direct logback usage required to overcome limitations of SLF4J MDC API. SLF4J MDC API heavily rely on [[ThreadLocal]], example: ch.qos.logback.classic.util.LogbackMDCAdapter Logback' [[LoggingEvent]] allow setting MDC directly as Java map that should have performance benefits compared with SLF4J/Logback implementation.
Please be aware that using other version of logback (than used in cats-helper-logback
) might bring '''RUNTIME ERRORS''' or '''MISSING LOGS''' in case of binary incompatibility between them.
Suggested approach is in using exactly same logback version as used in cats-helper-logback
(among all others available through transitive dependencies)
In some cases it may be necessary to allocate the logback instance manually as well as using the SLF4J API in the end user code. However, if multiple LoggerContexts are instantiated at the same time, this could lead to unexpected behaviour, such as the RollingFileAppender writing to multiple files instead of one.
To cover such cases, the internal implementation of LogOfFromLogback
uses the SLF4J API to instantiate the logback context, so that later use of the SLF4J API will pick up the same context instance created by LogOfFromLogback
.
This helper lives in a separate cats-helper-testkit
module. It is makes testing F[_]
-based code easier.
NOTE: cats-helper-testkit
is an experimental module and may break SemVer guarantees from time to time.
However we will do our best to avoid unnecessary breakages.
"what time is it now?" in PureTest[IO].of { env =>
import env._
for {
_ <- IO.sleep(1.hour)
_ <- testRuntime.getTimeSinceStart.map(_ shouldBe 1.hour)
} yield ()
}
addSbtPlugin("com.evolution" % "sbt-artifactory-plugin" % "0.0.2")
libraryDependencies += "com.evolutiongaming" %% "cats-helper" % "2.2.3"