April 2016


Integrated PR 171 which adds support for writing WebDriver specification tests in the Web Platform Tests repository.

Reviewed PR 178 to wptrunner which tries to deal with socket timeouts in Marionette.

Submitted PR 2745 to test the WebDriver navigator.webdriver IDL.

Reviewed bug 1243083 which enables e10s in Marionette by default.

Published the wptrunner WebDriver client as wdclient separately.

Submitted PR 62 to replace the WebDriver client in WPT with the one from wptrunner, now known as wdclient.

Provided info on bug 1261412.


Provided needinfo on bug 1235376 about context menu %s substitution in extensions.

Did some administrative duties with regards to conferences, &c.


Submitted PR 181 to wptrunner where the WebDriver client dependency gets picked up from wpt-tools.

Submitted PR 293 to the WebDriver specification to fix a case where the response result returns twice when an error occurs.

Submitted PR 2752 as the very first stab at writing WPT tests for the WebDriver specification. It adds tests for the chapter on the protocol.

Submitted PR 182 to support the is_alive check for unstarted processes of webdriver_server in wptrunner.


Fixed code review issue in r/6333.

1:1 with dburns.

Work on further WPT tests for WebDriver.

Sent an email to the Browser Tools- and Testing WG with a Window object serialisation proposal for the WebDriver specification.


Submitted PR 294 to the WebDriver specification which adds stale element checking to Switch To Frame as well as a couple of dependencies and references.

Also submitted PR 295 that provides an implementation for the proposal I sent yesterday regarding providing serialisations of WindowProxy objects.


Provided some input on bug 1262963.

Submitted PR 183 to wptrunner that ensures we provide wdspec tests with a pristine session. We do this by dismissing any open user prompts and by closing windows that had been opened by the test function.

Submitted PR 2799 to Web Platform Tests with some preliminary Get Current URL command tests for WebDriver.

Submitted PR 296 to define Get Element Text in WebDriver.

Submitted PR 297 to remove links to the multipage version of the HTML standard in the WebDriver specification.

Submitted PR 298 to improve the wording of the Element Retireval introduction in the WebDriver specification.


Reviewed bug 905650.


Addressed some code review issues in PR 64 to update Norwegian translations of

Disabled some CSP tests for bug 1123506 which we assume were incorrectly passing earlier. There seems to be a pattern that the test fail on debug and optimised e10s jobs. Perhaps the patch makes use of an e10s feature that is broken?

Filed bug 1263661 about associating user prompt dialogues with the current browser rather than the session in Marionette.


Updated test expectations for CSP tests related to bug 1123506, as explained yesterday.

Provided input on bug 1264104.


Submitted PR 57 to wires which silences a compiler warning about using private types in public interfaces.

Filed bug 1264259 about implementing the user prompt handler described in the WebDriver specification in Marionette.

Work on bug 1263661.


Attended office meeting. The AV people were invited.

Worked on 1263661 to associate user prompts with the browsing context instead of the global session state. It turns out that this.browser on GeckoDriver stores the outer window object, which again contains the tabbed browsers (the browsing contexts). I need to ask jgriffin about why the data structure has been designed this way. The issue is that you cannot set properties on this.curBrowser, because it is reset to the outer window BrowserObj on calling the Switch To Window command.










Followed up on PR 64 to the Mozilla website translations into Norwegian.

Commented on issue 2161 in WPT about implementing the meta-circular WebDriver client. Noted the concern on blocking.

Attended office meeting with Rob.




Provided needinfo request on bug 1266855 about the *.egg-info directory ignore rule I introduced in bug 1153822.

Tried to chase down bug 1123506 further with the help of bz. Rebased the patch and pushed it to try again as I will try to re-use the container it is run it to see if the test reproduces.


Rebased PR 297 about removing HTML multipage links from the WebDriver specification.

Requested needinfo from jryans on nsIServerSocket’s IPv6 capabilities for bug 1257719. It would be nice to support IPv6 connections in Marionette.

Rebased and address review issues for PR 296 about relying on innerText for the WebDriver specification.

Replied to review comments for my PR 295 to serialise Window objects in WebDriver.

Replied to the questions raised by Simon in my suggestion to provide Window object serialisation in WebDriver.


Reviewed PR 300 to the WebDriver standard.

Reviewed r/6402 to wptrunner.

Reopened bug 1247748.

Filed bug 1268049 about an index out of bounds error in mozharness when using the --try-test-paths flag with mach on try with Windows.

Worked on reproducing bug 1123506. Got one step further If you compare these two try runs you will see that it is in fact the CSP tests that are causing test_document_open.html to fail.


Filed bug 1268496 about triggering of incorrect test jobs across different platforms when using ./mach try with a test path.

Continued investigation into 1123506 with a try run that selects the CSP tests I updated the expectations for and the navigation timing tests succeeding them. It will be interesting to see if they are the cause of navigation timing’s test_document_open.html timing out.