2019

April

Monday 1st

Attended remote protocol meeting, thereafter prepared the minutes.

Filed bug 1540655 for running the Puppeteer test suite on try.

Commented on and closed WebDriver issue 1407.

Closed WebDriver issue 1266.

Some more Outreachy organisation.

Provided needinfo on bug 1483559.

Reviewed bug 1539221.

Filed and patched bug 1540684.

Reviewed bug 1539213.

Reviewed bug 1539210.

Reviewed bug 1538809.

Patched bug 1537775.

Patched bug 1537770.

Sent email about communication to remote protocol mailing list, following up on an action item from the cross-vendor Puppeteer/Juggler sync meeting we had 12 March.

Wednesday 3rd

More Outreachy organisation.

Dealt with geckodriver PR 1536.

Commented on geckodriver issue 1354.

Commented on geckodriver issue 1532.

Reviewed WPT PR 16225.

Commented on WPT PR 16222.

Reviewed bug 1539221.

Reviewed bug 1539210.

Addressed a review issue with bug 1537770.

Thursday 4th

Commented on bug 1483559 and filed bug 1541812. Requested two needinfos on the former bug.

Filed phlay issue 55.

Addressed further code review issues with bug 1537775.

Addressed code review issues with bug 1540684, then filed and patched bug 1541834 as a follow-up.

Work on bug 1533831.

Friday 5th

Updates to the Puppeteer roadmap.

Preparation for the remote protocol meeting on Monday.

Rebased bug 1541834.

Filed and patched bug 1542198.

Filed bug 1542229 to perform a security review of the remote agent.

Requested needinfo on bug 1542229.

Work on bug 1533831. It is not test_Domains.js that is missing, but a call to ChromeUtils.import() for remote/domains/Domains.jsm that fails.

Monday 8th

Attended remote protocol meeting. Afterwards I prepared the minutes.

Closed geckodriver issue 1539.

Commented on the review of bug 1525427.

Continued work on bug 1533831. I hate making changes to build system sensitive files.

Filed and patched bug 1542836.

Filed and patched bug 1542851.

Filed and patched bug 1542861 as a prerequisite to enable the remote agent on Nightly in bug 1533831.

Commented on bug 1542848.

Tuesday 9th

Closed bug 1542851.

Reviewed WebDriver PR 1410.

Commented on bug 1542848 regarding a possible regression caused by the remote agent.

Submitted a preliminary patch for bug 1526364 to introduce a visual cue when the remote agent is active.

Did some remote protocol bug triage.

Requested needinfo on bug 1535102.

Commented on bug 1543142.

Commented on geckodriver issue 1515.

Commented on geckodriver issue 1540.

Wednesday 10th

Filed bug 1543494.

Thursday 11th

Work on bug 1533831 to enable the remote debugger in Nightly.

Filed bug 1543630 to populate IMPACTED_TESTS throughout the remote agent.

Filed bug 1543676 and bug 1543679 to fix some Mochitests before the remote agent can be enabled by default.

Friday 12th

Reviewed bug 1543725.

Patched bug 1543676.

Patched bug 1543679.

Filed and patched bug 1543945.

Patched bug 1543946 and commented on geckodriver issue 1515.

Saturday 13th

Remote protocol meeting preparations.

Monday 15th

Attended remote protocol meeting.

1:1 with dburns.

Filed bug 1544393 to ship the remote agent in Nightly, which changes the narrative of bug 1533831 from shipping, to implying it in the default build but not enabling it.

Updated bug 1542229.

Addressed code review issue with bug 1543676.

Attended Puppeteer/Juggler sync meeting.

Submitted patches for bug 1533831 to enable the remote agent in the default Firefox build.

Tuesday 16th

Commented on geckodriver issue 1543.

Replied to code review comments for bug 1533831.

Rebased bug 1533831.

Added a commented on the review of bug 1544042.

Provided needinfo on bug 1495621.

Commented on bug 1543115.

Wednesday 17th

Reviewed D27784.

Filed puppeteer-mvp to track progress and communicate success towards shipping an MVP of the remote debugging protocol in Firefox.

Filed WebDriver issue 1412 to JSON serialise the return value of Get Element Property.

Rebased and addressed code review issues with bug 1533831.

Closed geckodriver issue 1542.

Commented on geckodriver issue 1543.

More work on bug 1533831 to get all the required tests to included on try.

Filed bug 1533831.

Filed bug 1545155 to have a remote protocol reviewers group created.

Commented on bug 1144075.

Tuesday 23rd

Commented on bug 1545724 with regards to beforeunload popup dialogues and dom.disable_beforeunload.

Filed and patched bug 1546303 which adds some basic tests for DOM property interaction with regards to WebDriver’s Execute Script and Execute Async Script commands.

Commented on bug 1544906, then filed and patched bug 1546319.

Commented on geckodriver issue 1543.

Commented on geckodriver issue 1546.

Closed geckodriver issue 1545.

Commented on geckodriver issue 1544.

Provided needinfo on bug 1533831 and filed—and patched—bug 1546385 as a follow-up to fix the failing browser_target.js test.

Closed bug 1429522.

Wednesday 24th

Provided needinfo on bug 1546610.

Rebased bug 1533831.

Filed and patched bug 1546714 to map some more remote protocol source files to bug components.

Filed and patched bug 1546722 to improve errors thrown in remote protocol browser-chrome tests by reconstructing them with RemoteAgentError.

Provided needinfo on bug 1533831.

Commented on WebDriver issue 1413.

Dealt with lots of backouts, fixups, and issues to make bug 1533831 stick.

Thursday 25th

Provided needinfo on bug 1543725.

Sent announcement to dev-remote that the remote agent is now included in default Firefox Nightly builds.

Commented on bug 1544906.

Filed and patched bug 1546933 and bug 1546935 to clarify some usage and test documentation, following bug 1533831.

Project planning for the remote protocol work.

Provided needinfo and addressed test failure that caused bug 1546714. to get backed out.

Addressed a code review issue with bug 1546722.

The patches to enable the remote agent in default Firefox Nightly builds caused browser-chrome tests on debug builds, where we run reference counting memory leak checks, to fail intermittently. I filed bug 1546945 to first disable the remote agent again to give us time to investigate. We later decided to re-enable it with the browser-chrome tests disabled on debug until we have time to investigate.

Friday 26th

Reviewed WebDriver PR 1415 and PR 1416.

Reviewed bug 1543095.

Reviewed bug 1543098.

Reviewed bug 1338040.

Commented on bug 1546945.

Commented on bug 1543115.

Wrote a proposal patch for a new chrome-only edump() function that lets you write to stderr from JS in bug 1543115. If that’s not acceptable, I have a backup plan for what can be done.

Monday 29th

Attended remote protocol meeting. Then prepared the meeting minutes after that.

Closed bug 1546850 and bug 1547303 as duplicates of bug 1546945, then filed and patched bug 1547679 to disable all the browser-chrome tests on debug until we have had time to investigate the memory leak in the remote agent.

Resigned as reviewer of bug 1544906.

Requested needinfo on bug 1540655.

Cleared needinfo on bug 1535102.

Filed and patched bug 1547700 to improve documentation on the remote agent browser-chrome tests, specifically regarding headless mode.

Tuedsay 30th

Commented on geckodriver issue 1540.

Submitted a follow-up patch for bug 1547679 to disable the remote protocol browser-chrome tests also on ASan.

Attended internal meeting about tracking milestone progress with the remote protocol.

Closed geckodriver issue 1540.

Attended internal team chit-chat.

Worked on a proof of concept for bug 1543115 that bootstraps the remote agent in a Rust XPCOM component. The idea is to implement nsICommandLineHandler in Rust, because it supports writing to stderr, and then to call nsIRemoteAgent implemented in JS from there. This has the benefit over the edump() patch that we would have to do this work regardless when we would make the switch from httpd.js to hyper.

Discussion with sole and jgraham about creating a dashboard for tracking CDP (Puppeteer?) progress.