set_multibuild_flavor is now setting only the multibuild container
as the method name implies.

Doing so no matter if the package comes via a link or not, since it does
not matter for container name.

do not abort due to wrong guessed flavors because the wrong source state might be
read.

options[:package] can not be a string for releasing as it would crash
when trying to lookup the release name, so having a seperate string
for container name. (and to avoid the logic that it is a multibuild, if
it is not a package object)

simplify authorization object code

Create Workflow::Step::BranchPackageStep model

[backend] support DoD container registries

Those DoD resources are different to repositories because there
is no way to request the complete state from the server.

Thus we collect the needed DoD resources in the scheduler and put
them into a "doddata.needed" file. The bs_dodup server then uses
this information to request data for each resource.

- check the blobstore before downloading a blob
- put downloaded blobs in the blobstore
- try to get the container mtime from the config so that
  we always reconstruct bitwise identical containers
- add correct provides to the obsbinlnk file

Remove package id from event table

The error messages are not used anywhere, therefore no need
to add them for now.

Co-authored-by: Daniel Donisa <daniel.donisa@suse.com>
Co-authored-by: Saray Cabrera Padrón <scabrerapadron@suse.de>

Co-authored-by: Daniel Donisa <daniel.donisa@suse.com>
Co-authored-by: Eduardo Navarro <enavarro@suse.com>
Co-authored-by: Saray Cabrera Padrón <scabrerapadron@suse.de>

After changing the from the attempt of building an association
between an event and a package, to fetching the package
using the passed payload in the `ReportToScmJob`
in https://github.com/openSUSE/open-build-service/pull/11124
the package_id is not longer needed.

Preperation for the db migration in order to remove the package_id
column from the events table.
See https://github.com/ankane/strong_migrations#removing-a-column

Hide scm channel from subscription configuration page

The test needs to take into account the hidden channels

The scm channel in the event system is not meant to be
set by hand, its only used internally to report the state of a
workflow (at least for now). Therefore we should not allow
to set it by hand (since it doesnt bring any value).

Create TriggerWorkflowController

Refine validations related to SCM webhooks in trigger controller

Add systemd service for scm delayed job queue

[src/api] Update puma: 5.3.1 → 5.3.2 (patch)

[src/api] Update ffi: 1.15.0 → 1.15.1 (patch)

Skip validation of params only if request comes from and supported SCM

Co-authored-by: Daniel Donisa <daniel.donisa@suse.com>
Co-authored-by: Dany Marcoux <dmarcoux@suse.com>
Co-authored-by: Eduardo Navarro <enavarro@suse.com>
Co-authored-by: Saray Cabrera Padrón <scabrerapadron@suse.de>

Only users with the `trigger_workflow` feature flag enabled can create
and trigger a workflow token.

Co-authored-by: Eduardo Navarro <enavarro@suse.com>

Return a "Bad request" error when the HTTP request headers are not
valid, for the trigger/workflow endpoint.  The response returned
previously was a "Forbidden" error, that in this case makes little
sense.

Return a BadScmPayload error for empty or wrong JSON payload in the
/trigger/workflow endpoint.

Replace `require_login` with Pundit in Webui::Staging::ProjectsController

[src/api] Update nokogiri: 1.11.4 → 1.11.5 (patch)

Add a remote boolean for Distribution

[src/api] Update flipper: 0.20.4 → 0.21.0 (major)

Revert "Apply the background color all the way to the right"

... to the right" and "Only create a CSS grid on code inside diffs".

This reverts commit 55d02ddc and commit
7d564eb9.

Creating a work around with CSS grid to apply a background color all the
way to the right breaks diffs with more than 1000 lines, both in Chrome
and Chromium. And breaks diffs with more than 10000 lines in Firefox.

This solution could be used for small diffs, but not for all of them.

Avoid sorting Actions DataTable's column

In preparation to move remote Distributions from memcache to a
database cache so we can use ActiveRecord for them.

The `ReportToScmJob` (responsible for reporting back the resulting
build states of a Workflow to GitHub/GitLab) queues the jobs
in a dedicated delayedjob queue.

Co-authored-by: Dany Marcoux <dmarcoux@suse.com>
Co-authored-by: Victor Pereira <vpereira@suse.com>

Fix data migration timeline dependency