December 2015


Provided more information on bug 1228079, but requested more info from a sheriff on the question of whether this should apply to all trees.

Discussed eslint with dburns.

Landed bug 1154432 on inbound.

Provided input on various eslint issues in bug 1228761 and bug 1229194.

Worked on a rewrite of the cookies infrastructure in WebDriver again.


Rebased bug 1211489 and triggered a new try run. Pushed to inbound again.

Filed bug 1229765 about monitoring the browser process Marionette spawns for crashes, following dbaron’s reported bug 1229549 yesterday.

Provided some input on bug 1229578 about testing that crashes and memory leaks are handled responsibly in test harnesses.

1:1 with dburns.

Submitted PR 282 to WebDriver which rewrites the cookies infrastructure to use concepts borrowed from RFC 6265.

Resolved bug 1217598 as invalid due to the previous PR to WebDriver mandating the input type of the expiry field to be an integer describing the number of seconds since the Unix Epoch.


Did some digging on bug 1230079 filed by the sheriffs after the b-m job started failing following bug 1211489 landed on inbound. Impressive triaging by nigelb.

Filed bug 1230151 as it turns out the Python client does not enforce the message sequencing introduced by bug 1211489 earlier this week. There is also a bug in the test decorator that causes it to always ignore the test.

Explained that bug 1230269 is due to an old Marionette client that is not compatible with protocol level 3 introduced by bug 1211489.



Submitted patch for bug 1230151.

Resolved bug 29321 for the WebDriver specification.


Travel day to Orlando.

Responded with feedback to W3C bug 29278.

Feedback on bug 1213875 regarding scroll behaviour for the Take Element Screenshot command.

Submitted PR 283 to WebDriver about returning the window dimensions when invoking the Set Window Size command.


Work week in Orlando.

Provided more info on bug 1164078.

Provided feedback for bug 1230847 about broken parallel emulator callbacks.

Provided feedback on W3C bug 29316 about how the Back command navigates the top-level browsing context rather than the current browsing context.


Work week in Orlando.

Reviewed bug 1230847 which fixes parallel emulator callbacks following the integration of bug 1211489.

Made a redaction to my comment of W3C bug 29316 about the behaviour related to the Back command in WebDriver.

Filed bug 1231423 about logging the protocol version in Marionette.

Reviewed bug 1231556 which removes some dead code from marionette-transport.


Work week in Orlando.

Attended build system planning meeting.

Helped out in the Firefox OS Emulator work session.


Work week in Orlando.


Travel day.


Still traveling…


Investigated bug 1230151 with jgriffin where Ld jobs are failing mysteriously. The bug is regarding enabling the skipping of out-of-sync messages in the Marionette Python client.

Some minor work on bug 1123506 to enable evaluation of content scripts with lasting side-effects.

Provided a bottle of line aquavit to colleagues in the London office. Wrote about its history to the office mailing list.


Brought some clarity to bug 1175116 about the missing stacktraces in the Marionette Node.js client.

Work on bug 1123506 to evaluate scripts in content with lastin side-effects. Were able to get primitive emulator callbacks in content working with a new promise-based infrastructure.



Conducted some maintenance work on, amongst things I rewrote the xentries program to Go.


Investigation into why the Ld job for bug 1230151, which enables message skipping in the Python client. Triggered new try job.




Worked on ironing out the last bugs of the Mn job from the try runs of bug 1123506 about evaluating scripts in content with lasting side-effects. Triggered try run to see if my fixes helped.

Provided an update on the progress of the work so far and the problems I am seeing in the latest try run.

Took some time to write two, what I think are, useful contributions to the dev-fxos@ mailing list.

Filed, and hopefully patched, bug 1234063. This sends associates a sandbox parameter to all invocations of executeScript, executeAsyncScript, and executeJSScript. This should allow me to integrate the PoC that is found in bug 1123506.


Replied to bug 1212609 with suggestions on where I think it’s best to place the Firefox UI tests. I think putting them inside testing/marionette makes little sense because it gets the dependency chain wrong: The Marionette implementation itself does not rely or depend on these harnesses or tests. On the contrary, they are consumers of the Marionette server implementation.

Fixed up some peer review issues with bug 1234063.

Things I need to do next for bug 1123506:


Reviewed bug 1212609 which moves the Firefox UI tests and other dependencies in-tree.

Provided input on bug 1208725 about increasing the B2G emulator timeout.

Filed bug 1234612 about logging where the virtualenv was attempted created.


Filed bug 1235365 about about:addons double encoding unicode characters.

Filed bug 1235376 about the %s format specified not being replaced for WebExtension context menu items with the context type "selection".


Provided instructions on bug 1198797.

Filed bug 1235553 and bug 1235554 about a missing histogram in Telemetry causing a lot of errors in the browser console.

Improved the MDN article on running Marionette tests.


Reviewed bug 1198797.

Provided info on bug 1235554.

Filed bug 1235766 about Firefox treating WebExtensions with the backslash character in the extension description as corrupt.

Followed up on bug 1235365.

Provided info on bug 1210778 about how to add debug statements to the Marionette server.

Redirected patch for bug 1097676 to ahal.

Submitted patch for bug 1234612 to include the path of the virtualenv we fail to create. This spun out of some annoyance debuggin this issue for moz-js/Servo.

Confirmed bug 1231871 about the lack of profile support when using Marionette with the Selenium Python bindings.