1144869 - Add BrowserTestUtils.withNewTab

Status: RESOLVED FIXED

(Testing :: General, defect)

Description

[:Paolo Amadini]

This is an utility function to execute a test on a web page.

Comment 1

[:Paolo Amadini]

Attachment: The patch

The first user would be the login tests:

yield BrowserTestUtils.withNewTab({
  gBrowser,
  url: "https://example.com/browser/toolkit/components/" +
       "passwordmgr/test/browser/form_basic.html",
}, function* (browser) {
  // ...
  yield ContentTask.spawn(browser, null, function* () { /*...*/ });
});

Comment 2

[Steven MacLeod [:smacleod]]

Comment on attachment 8579604 [details] [diff] [review]
The patch

Review of attachment 8579604 [details] [diff] [review]:
-----------------------------------------------------------------

::: testing/mochitest/BrowserTestUtils/BrowserTestUtils.jsm
@@ +52,5 @@
> +   */
> +  withNewTab: Task.async(function* (options, taskFn) {
> +    // In theory we may be subject to a race condition here, since we cannot set
> +    // up the browserLoaded listener before calling addTab. In practice this is
> +    // how most tests currently do their setup, so this is expected to work.

So I believe this is guaranteed to work because the way we're listening for the load is in the content-utils.js framescript, and then sending an async message up. Since we don't yield the event loop before setting up our listener for that message (in browserLoaded) we can't miss it.

That being said I don't think we need this comment about a possible race condition. What I said above should probably be documented in browserLoaded though.

If you still think this comment is needed, I'm fine with that too though.

Comment 3

[:Paolo Amadini]

(In reply to Steven MacLeod [:smacleod] from comment #2)
> So I believe this is guaranteed to work because the way we're listening for
> the load is in the content-utils.js framescript, and then sending an async
> message up. Since we don't yield the event loop before setting up our
> listener for that message (in browserLoaded) we can't miss it.

You're right, I've documented this as you suggested and pushed to fx-team:

https://hg.mozilla.org/integration/fx-team/rev/477656ff3fc5

Comment 4

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/477656ff3fc5