Love Weaver? Like Playwright? Forced to write tests?

Very small integration between Playwright and Weaver to facilitate browser testing of applications.

Here's an example of usage which is also a Scala CLI script!

test.scala

//> using lib "com.indoorvivants.playwright::weaver:0.0.4"
//> using lib "com.disneystreaming::weaver-cats:0.8.0"

import com.indoorvivants.weaver.playwright._
import cats.effect._

object Example extends weaver.IOSuite with PlaywrightSpec {
  def sharedResource: Resource[IO, Res] = PlaywrightRuntime.create()

  pageTest("hello playwright!") { pc =>
    for {
      _ <- pc.page(_.navigate("https://playwright.dev"))

      _ <- pc.locator("text=Get Started").map(_.first().click())

      _ <- eventually(pc.page(_.url())) { url =>
        expect(url.contains("intro"))
      }
    } yield success
  }
}

Run it with scala-cli test test.scala.

For more common build tools:

SBT

"com.indoorvivants.weaver" %% "playwright" % "<version>" % Test

Mill

ivy"com.indoorvivants.weaver::playwright:<version>"

1. Wraps Playwright and browser instantiation into a Resource

   • And also handles bonkers exceptions it throws when the drivers are created concurrently

2. Does minimal wrapping of the side-effectful, non-RT APIs that Playwright exposes

   • As I learn more about frontend testing, the number of wrapped APIs should increase, but the goal is to mainly access them via the .page(...) combinator

3. Puts semaphores in place to abide by Playwright's recommendation

4. Provides a eventually combinator for retrying assertions

As you can see, it's intentionally minimal