Bug 1614763 - [ci] Create beetmover tasks to publish release artifacts
Added TODO
Added head_tag to try_task_config
Change locale to multi and fix beetmover URL destinations
Bump version
Adjust beetmover kind and add beta + format task label
Allow all build types for beetmover
Bump version
Adjust beetmover destination URL + minor mods
Try task config - nightly
Change try release to beta
Remove try_task_config
We've noticed that geckoview_example is ~300ms faster than fenix in cold page
load tests on arewefastyet for the Pixel 2. We suspect the main
difference is because geckoview_example runs with conditioned profiles
and fenix does not.
This PR is foremost an experiment to see if that's true because, after bug
1587542, we cannot run get results for fenix perftest PRs (i.e. it needs to be
merged into main). If we find that the results are not noisy, however, we
could end up leaving this in the tree. We've previously seen excessive
noise with fenix start up tests with conditioned profiles which is why
conditioned profiles are not currently enabled.
This was moved to once per day in https://bugzilla.mozilla.org/show_bug.cgi?id=1628413 when Google Play reviews were very slow. That hasn't been a problem for awhile, and the less frequent nightlies are causing delays getting new Android Components bumps shipped out.
* Fixes#11427 - Rename master branch to main (Automation/TC)
* Fixes#11427 - Rename master branch to main (GitHub Workflows)
* Fixes#11427 - Rename master branch to main (GitHub Workflows)
* Fixes#11427 - Rename master branch to main (GitHub Issue Templates)
* Fixes#11427 - Rename master branch to main (Jenkins)
* Fixes#11427 - Rename master branch to main (README & Documentation)
* Fixes#11427 - Rename master branch to main (Jenkins)
* Fixes#11427 - Rename master branch to main (Random)
* For #19801 - Update to latest version of taskgraph
* For #19801 - Update to latest decision-mobile image
We need to explicitly set the Java version as Java 11 now co-exists in
the decision-mobile image.
* For #19801 - Run decision task with Python 3
Instead of using a timestamp this will synchronize the Nightly version with the version
of Gecko / A-C. So when using 90.0.20210426143115 then the Nightly version will be
90.0a1 - similar to what Firefox desktop uses.
* run-ui-test-nightly
* Run ui tests on nightly build
* set correct number of shard according to tests running
* fix typo
* removing testing file
Co-authored-by: isabel rios <isabelrios@MBP-de-isabel-2.home>
Let's make mozilla-releng/staging-fenix the official staging repository.
By supporting it in the automation, we no longer need to maintain a
staging patchset to test things properly.
This PR builds on [AC##9024][1], and implements setting the Nimbus endpoints from [a secret set at build time][2].
For production use, this requires a secret named `nimbus_url` to be put into CI.
Note: Nimbus is currently behind a feature flag.
If developers wish to use a Nimbus server for local development, you can set the url by adding an entry into local.properties, e.g.:
```
nimbus.remote-settings.url=https://settings.stage.moz4ws.net
```
Without setting server, Nimbus will be able to function, except no experimental definitions will be fetched, and features under experiment will be configured as if not enrolled in the experiment.
[1]: https://github.com/mozilla-mobile/android-components/pull/9024
[2]: https://groups.google.com/a/mozilla.com/g/android-components-team/c/lAGVKQy8aiA/m/rY3uGAwhBAAJ
* Enable webrender for all tests and run a subset without webrender.
* Run tests on PR.
* Change task labels for webrender-enabled tests.
* Move transform logic to after the main transform.
* Reformat group symbol.
* Check if extra field is empty.
* Try a different method for treeherder info.
* Fix up assignment issue.
* Reformat symbol field instead of groupSymbol.
* Add new task group to config.
* Change the platform name for webrender tasks.
* Undo testing changes.
* Undo platform naming changes.
`assemble` doesn't assemble the tests so we need to run `test`.
`testClasses` isn't good enough because, according to `--profile`, it
doesn't include dependencies for `testRuntimeClasspath`.
Currently, push-apk fails to verify its chain of trust if github tagging
is done before it runs. Until this is fixed, we need to make sure
tagging is blocked on pushing.