Currently the default logging is set to debug in dev mode. This causes a LOT of logging output that can get cumbersome to develop with.

The best solution I have found is to do something like this:

config :logger,
  backends: [:console],
  compile_time_purge_matching: [
    [application: :playwright]
  ]

This requires recompiling the dependency: mix deps.compile playwright --force. This may also require any application that uses this to restart as well.

The issue here is there is no way to turn logging on for a portion of the use of playwright-elixir and then off.

I've seen FlexLogger but I dont want to add additional dependencies.

Any ideas?

Hi!

I managed to get the Phoenix sandbox set up, and wrote about it on my blog. I found that the existing PlaywrightTest.Case module, the Playwright.Browser.new_context function and Playwright.Browser.new_page were not sufficient for my needs so I had to copy and modify them. Maybe this can be improved by changing the library?

As a note, I think it may be better to configure the user-agent header instead of a separate one like I did, but then I think the metadata should be appended to existing user agent that Playwright would use, and I couldn't find a way to read it. Maybe there's a way, or maybe such feature can be requested from upstream. Ideally the user could do either one, by making the library flexible and allowing to configure both the user-agent and additional headers.

example:

test/playwright_test.exs:7
** (FunctionClauseError) no function clause matching in Playwright.Browser.new_page/2

The following arguments were given to Playwright.Browser.new_page/2:
    # 1
    {:error, %Playwright.Channel.Error{message: "Timeout 30000ms exceeded."}}
    # 2
    %{}
  
Attempted function clauses (showing 1 out of 1):
  def new_page(%Playwright.Browser{session: session} = browser, options)

stacktrace:
  (playwright 0.1.17-preview-6) lib/playwright/browser.ex:153: Playwright.Browser.new_page/2
  test/playwright_test.exs:3: Playwright.PlaywrightTest.__ex_unit_setup_0/1
  test/playwright_test.exs:1: Playwright.PlaywrightTest.__ex_unit__/2

In the response event, we now have response |> Playwright.Response.body() which works great when the content type is "image/gif" or similar however when the response type is "application/json" or "application/json; charset=utf-8" it seems to hang for me. Not sure if this is something to do with gzip encoding the json content-type. Any recommendations would be helpful.

Hi 👋 ! Thank you for this.

I'm just getting my feet wet with Playwright and your library, so maybe I'm doing something wrong. I noticed that Playwright has an ability to fill an eventhandle, but I do not see that this is exposed by this library.

If there's another way to do what I need to do, please let me know. I'm trying to work within an iframe. I can get the iframe and get the eventhandle from a query and then focus it. But I need to then fill it.

A simple change to PLAYWRIGHT_RUN_ASSET_SERVER should suffice to switch modes.

Is there a way to retrieve listener events for page requests? I see the Page.on/3 and I see it can receive console type events, but what about Request events as per Playwright?

How can I listen to requests?

Hello,
I would like to ask you about implementation abort function so we can skip some resources to load, especially in my case images and videos.
I made this function here:
https://github.com/geometerio/playwright-elixir/blob/ab729b6f6f1f4f6d9e238d5e97598757b968c548/lib/playwright/route.ex#L14-L15

@spec abort(t(), binary()) :: :ok
  defp abort(%Route{session: session} = route, error_code \\ nil) do
    Channel.post(session, {:guid, route.guid}, :abort, %{error_ode: error_code})
  end

and implementation:

Playwright.Page.route(page_loader, "**/*", fn route, request ->
      if request.resource_type in ["image", "font"] do
        Playwright.Route.abort(route, "blockedbyclient")
      else
        Playwright.Route.continue(route)
      end
    end)

But it crash server, I very new in playwright so not sure where can be problem.

Doc for abort is here and my code is here

For real-time communication use cases, it would be nice to be able to write a single test case that utilizes and exercises multiple browser types.

This repository is going to move from the geometerio organization to mechanical-orchard, where we intend to contribute more actively.

I believe that this is intended to be the executable_path from LaunchOptions.

@dbrody, thoughts on this?

cc @woodward

Something like the following occurs, infrequently. Once this timeout triggers, all following tests fail. This behavior appears to be non-deterministic: mix test --seed <seed> (where <seed> is that from a previous failure) does not force the failure.

  1) test Page .query_selector_all/2 (Test.Features.PageTest)
     test/features/page_test.exs:27
     ** (ExUnit.TimeoutError) test timed out after 60000ms. You can change the timeout:
     
       1. per test by setting "@tag timeout: x" (accepts :infinity)
       2. per test module by setting "@moduletag timeout: x" (accepts :infinity)
       3. globally via "ExUnit.start(timeout: x)" configuration
       4. by running "mix test --timeout x" which sets timeout
       5. or by running "mix test --trace" which sets timeout to infinity
          (useful when using IEx.pry/0)
     
     where "x" is the timeout given as integer in milliseconds (defaults to 60_000).
     
     code: [outer, inner] = Playwright.Page.query_selector_all(page, "css=div")
     stacktrace:
       (stdlib 3.16.1) timer.erl:152: :timer.sleep/1
       (playwright 0.1.1-preview) lib/playwright/page.ex:289: Playwright.Page.hydrate/1
       (elixir 1.12.3) lib/enum.ex:1582: Enum."-map/2-lists^map/1-0-"/2
       test/features/page_test.exs:30: (test)
       (ex_unit 1.12.3) lib/ex_unit/runner.ex:502: ExUnit.Runner.exec_test/1
       (stdlib 3.16.1) timer.erl:166: :timer.tc/1
       (ex_unit 1.12.3) lib/ex_unit/runner.ex:453: anonymous fn/4 in ExUnit.Runner.spawn_test_monitor/4

When trying to use playwright-elixir in an existing Phoenix project, I get the following dependency error:

Failed to use "gun" (version 2.0.0-rc.2) because
  playwright (version 0.1.16-preview-2) requires ~> 2.0.0-rc.2
  tesla (version 1.4.3) requires ~> 1.3

Gun hasn't had any commits since this RC was cut: https://github.com/ninenines/gun

We should evaluate if we really need to be on this version of Gun

A trace example:

** (FunctionClauseError) no function clause matching in Playwright.Browser.new_page/2
    (playwright 1.18.0-alpha.1) lib/playwright/browser.ex:153: Playwright.Browser.new_page({:error, %Playwright.Channel.Error{message: "Executable doesn't exist at <redacted>/chrome-mac/Chromium.app/Contents/MacOS/Chromium"}}, %{})
    <redacted>
    (elixir 1.13.2) lib/task/supervised.ex:89: Task.Supervised.invoke_mfa/2
    (elixir 1.13.2) lib/task/supervised.ex:34: Task.Supervised.reply/4
    (stdlib 3.17) proc_lib.erl:226: :proc_lib.init_p_do_apply/3

Before I dive in and try to push Keyboard features, anybody else work on this? I'd hate to duplicate effort. Also, if theres any pointers or suggestions, this would be a good way to track it.

Currently the getting started page is blank and the examples are all in test cases. I think improving the documentation could help adoption and use of the package.

Hi!

I saw that other similar projects (playwright-sharp and others) generate the client modules from playwright API specification JSON file.

Is there plan to do the same here? There are multiple ways to do that in Elixir I think, one could be an approach similar to what chrome_remote_interface Elixir package does.

This adds an extra integration step to every application that tries to use this library.

Ideas:

• Add documentation about this dependency and how to resolve it.
• playwright-elixir can ship playwright in its priv directory.
• This library can check for playwright being installed in the environment, and tell you how to install it.

Some work has already been done to extract the test assets to playwright-assets. It would be nice to have that be capable of loading and running as a dependency for this project so that:

• playwright-assets is usable for other projects. These should not necessarily have to be Elixir projects. So, a standalone server is needed.
• playwright-assets is able to be independently updated and version-referenced.
• playwright-elixir is not reported by GitHub to be a project made up mostly of Javascript.

In keeping with Elixir conventions, it would be nice to have extensions to the typical Playwright function that raise errors in certain cases, rather than returning nil, for example.

When opening browser via Playwright.BrowserType.launch(:chromium, options) and then closing the browser via Playwright.Browser.close(), the port opened to run chromium seems to stay open.

Seems to stem from within Playwright.Transport when called with Transport.Driver here:
https://github.com/geometerio/playwright-elixir/blob/17abee407e2d5fa30cc42ce62e8017c3d0ae8680/lib/playwright/transport/driver.ex#L18-L22

Even though all the GenServers are started with start_link it doesnt seem to clean up the port.

That's a concern internal to playwright-elixir.

In my setup I'm seeing the performance of calls to playwright elixir take longer and longer the more actions are performed. I think it may have to do with callbacks or something? I do have a listener on responses.

It may be my setup but I'm curious if others see this or the best approach to test this issue?