Squad

This 1.14 release adds an API endpoint that allow comparison
between builds from the same project. It also fixes a bug related
to LAVA ci backends when retrieving listener url.

The release also improves the docker image and pushes it to
https://hub.docker.com/repository/docker/squadproject/squad
on every release.

Complete list of changes going in:

* api:
* rest: endpoint to compare two builds
* rest: fix environment filter in knownissue
* ci:
* fix LAVA listener port retrieval
* fix get_listener_url() in case of REST
* frontend: comparison: fix transitions
* misc:
* Dockerfile: improve Dockerfile
* release: release docker image along with pip release
* remove CODEOWNERS file
* requirements: fix dependencies for broken celery 4.4.4
* settings: allow SMTP port customization
* test-docker: remove extra testing under docker
* travis: add verbosiness do test under docker

This 1.13 release marks 4 years of SQUAD project, yay! It also
adds a logo for the project along with other small bug fixes and
new features.

The LAVA backend is now compatible with LAVA RESTfull interface, also
now we're running tests against a real version of LAVA running on a Docker
container.

There was a small change in the workflow to check test details. The testrun
page has been completely removed in favor of less clicking to get to test metadata
and history.

The linux log parser plugin now captures a new class of bug called "invalid opcode",
and it catches more of "BUG" bugs.

Complete list of changes going in:

* ci:
* Enable REST API in LAVA backend
* fix ProjectStatus processing in fetch()
* improve TestJob.cancel()
* test lava backend using real LAVA instance.
* core:
* Implementation of ProjectStatus baseline
* add ProjectStatus baseline field
* add compute_project_statuses command
* fix compute_build_summaries command
* frontend:
* build: fix missing "failures only" filter
* add test details and remove testrun page.
* plugins: linux_log_parser: improve BUG matches and add "invalid opcode"
* misc:
* add logo to the README
* add logo to the main template.
* update the copyright years line to extend to 2020

This 1.12 release adds a bell icon to project UI allowing users
to subscribe/unsubscribe from projects notifications, among other
small fixes and additions.

Complete list of changes going in:

* api: add docstring to endpoints with additional methods
* api: rest: handle null test_name in project/test_results
* frontend: add subscribe feature to project listing
* frontend: allow to subscribe/unsubscribe from project view
* static: annotation: better handle error messages
* test: fix flake8 warning

This 1.11 release adds several new features. A new flag in Project/Backend
settings allow processing valid results of LAVA jobs that had an Infrastructure
error. In the test job page of a build, a user can now choose to cancel a job that
is submitted but not yet fetched, also sending this cancel action to job backend.

Complete list of changes going in:

* api:
* document field lookups
* rest: add ID to all possible serializers and filters
* rest: add resubmit and cancel methods to TestJob objects
* ci:
* add support for cancelling CI jobs
* lava: add flag allowing to accept results from failed jobs
* models: handle race condition without wait
* doc:
* api: add note about Squad-Client tool
* api: add status to testrun endpoint and metrics endpoint
* api: remove usergroups documentation
* intro: add a note about xfail tests
* intro: add note on regressions and fixes
* frontend:
* add Cancel button to test job list
* add link to full log from test list page
* ci: reduce number of queries for testjobs view
* core: fix project counter display on index page

This is a bug fix release:

In case the job is still in the queue or is in progress, fetch() raises
TemporaryFetchIssue from within block that catches FetchIssue. Since
TemporaryFetchIssue derives from FetchIssue, the exception is
immediately consumed and fetch_attempts counter is increased. This leads
to maxing out fetch_counter for jobs that stay in the queue for a long
time. This patch removes the exception and instead returns from the
fetch() function without an error. This doesn't result in increasing
fetch_attempt counter.

Complete list of changes going in:

* api: rest: add ID field to TestJob object
* ci: fix Backend.fetch() implementation
* travis: fix reuse docker lint
* Translated using Weblate (French)

This 1.10 release fixes a race condition bug that happens whenever
two or more workers fetch a single TestJob, triggering a fetch issue.
Also it adds suport to Sentry, a tool to manage log events.

The release also create a separate endpoint for Metrics objects.

Complete list of changes going in:

* api: rest: fix force-created reports
* ci: models: fix race condition for fetch
* core:
* rest: add top-level metric endpoint
* tasks: fix job_id as integer
* tasks: notification: keep only one task for notification on timeout
* tasks: remove try-catch block of notification task
* manage: exclude tests that break on sqlite
* settings: add support to Sentry