Engineering Effectiveness Newsletter (February 2022 Edition)

Welcome to the February edition of the Engineering Effectiveness Newsletter! The Engineering Effectiveness org makes it easy to develop, test and release Mozilla software at scale. See below for some highlights, then read on for more detailed info!

Highlights

Detailed Project Updates

Bugzilla and Bugbug

Build System and Mach Environment

CI and Treeherder

  • arai added a button to task groups which offers a one-click solution to request all tasks in a group to run - commonly used by developers testing their changes on Try to extend the build and tests to run (e.g. empty pushes)
  • Eva added support for retrieving automated mozci failure classification results and store them in Treeherder
  • marco landed multiple improvements to the mozci intermittent / real failure classification logic (1, 2, 3, 4, 5)
  • Eva implemented a monitoring mode to periodically evaluate mozci results and investigate errors
  • Bastien added support for sending emails when mozci detects a push containing a real failure, and Eva for sending Matrix notifications
  • Eva built a script to compare mozci classifications with sheriff classifications (1, 2)
  • Marco and Eva improved the mozci classification evaluation script to avoid any biases from manual classifications and expanded it (1, 2, 3, 4, 5)
  • jmaher got all desktop tests running as fission, and added a Linux “nofis” variant until Android is supported
  • jmaher enabled artifact support for browser-chrome a11y tests, split up the large mochitest-media manifests, and “greened up” CI by improving the way some intermittent tests were run
  • Chidinma added the ability to retrigger tasks from mozci as well as extended the classify command to auto-retrigger for greater accuracy

Fuzzing and Sanitizers

  • Jesse added fuzzing support of Mac OS on Taskcluster
  • Jason implemented WebGPU fuzzing in Domino

Lint, Static Analysis and Code Coverage

PDF.js

Phabricator , moz-phab, and Lando

  • sheehan landed changes to Lando to allow testing against a mock s3 bucket, eliminating the need for a real one

Release Engineering and Management

  • ahal created comprehensive documentation on getting started with Taskgraph for the first time.
  • ahal designed and implemented a way to install Python dependencies from the run-task script. This will allow consumers of Taskgraph to install it from pypi.
  • hneiva refactored VPN signing and added support for adhoc signing. This will enable automated signing once builds are moved to Taskcluster.
  • Gabriel enhanced the add-ons pipeline to archive and create releases for signed system add-on artifacts.

Version Control

  • sheehan added a check for the Python version to the Mercurial setup wizard, to prepare for Mercurial Py2 deprecation

Other

Thanks for reading and see you next month!