Add settings for minimum key strength and allowed key type (EE)

What does this MR do?

EE version of https://gitlab.com/gitlab-org/gitlab-ce/issues/17849 https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/13712

assigned to @rspeicher

assigned to @nick.thomas

Some good spec failures here

OK, these should pass now.

assigned to @rspeicher

@nick.thomas Pending CE changes.

assigned to @nick.thomas

restored source branch ee-version-of-17849

added 479 commits

• 5619027f...b74c662e - 477 commits from branch gitlab-org:master
• 2af46841 - Add settings for minimum key strength and allowed key type (EE)
• 5fbf0440 - Address review comments

Compare with previous version

This now matches https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/13712 but has one seemingly genuine spec failure:

https://gitlab.com/nick.thomas/gitlab-ee/-/jobs/31075293

I guess it's in the CE->EE differences, I'll dig.

OK, I expect this to pass now. We were using key.key in the spec rather than key.key_text.

@rspeicher can you review?

assigned to @rspeicher

mentioned in issue #3310

resolved all discussions

approved this merge request

changed milestone to %10.0

added Community Contribution Platform labels

Thanks Nick!

@rspeicher those two specs pass locally :/ https://gitlab.com/nick.thomas/gitlab-ee/-/jobs/31153264

I suspect an order dependency. Should we try to fix it before merging? I'm a little worried about monday's CE->EE merge if we don't get this in.

added 8 commits

• 201d6160...71f09bb7 - 6 commits from branch gitlab-org:master
• d40a7d41 - Add settings for minimum key strength and allowed key type (EE)
• 5a0c0e15 - Address review comments

Compare with previous version

I've tried rebasing against latest master, to see if that helps.

It didn't. Order dependency suspected, so I bisected:

lupine@gitlab-t470p:~/dev/gitlab.com/gitlab-org/gdk-ee/gitlab$ bundle exec rspec --bisect spec/features/projects/audit_events_spec.rb spec/features/groups/audit_events_spec.rb spec/features/promotion_spec.rb
Bisect started using options: "spec/features/projects/audit_events_spec.rb spec/features/groups/audit_events_spec.rb spec/features/promotion_spec.rb"
Running suite to find failures... (3 minutes 12.8 seconds)
Starting bisect with 2 failing examples and 27 non-failing examples.
Checking that failure(s) are order-dependent... failure appears to be order-dependent

Round 1: bisecting over non-failing examples 1-27 .. ignoring examples 15-27 (3 minutes 39.2 seconds)
Round 2: bisecting over non-failing examples 1-14 .. ignoring examples 8-14 (2 minutes 51.3 seconds)
Round 3: bisecting over non-failing examples 1-7 .. ignoring examples 5-7 (2 minutes 30.2 seconds)
Round 4: bisecting over non-failing examples 1-4 .. ignoring examples 3-4 (2 minutes 12.1 seconds)
Round 5: bisecting over non-failing examples 1-2 . ignoring example 1 (1 minute 32.17 seconds)
Bisect complete! Reduced necessary non-failing examples from 27 to 1 in 13 minutes 10 seconds.

The minimal reproduction command is:
  rspec ./spec/features/projects/audit_events_spec.rb[1:1:1] ./spec/features/promotion_spec.rb[1:10:1,1:11:1]

Since the specs are failing due to 404, I expect that the audit_events_spec is disabling audit events in the license somehow. Looking.

OK, the issue is that @show_promotions in the LicenseHelper module is being cached between specs. Some puts debugging reveals:

$ bundle exec rspec -f d ./spec/features/projects/audit_events_spec.rb[1:1:1] ./spec/features/promotion_spec.rb[1:10:1]
Projects > Audit Events
Starting the Capybara driver server...

== Seed from /home/lupine/dev/gitlab.com/gitlab-org/gdk-ee/gitlab/db/fixtures/test/01_plan.rb
  unlicensed
object_id: 46970522063440 self: LicenseHelper
    returns 404

Promotions

== Seed from /home/lupine/dev/gitlab.com/gitlab-org/gdk-ee/gitlab/db/fixtures/test/01_plan.rb
  for project audit events
object_id: 46970522063440 self: LicenseHelper
    should appear on the page (FAILED - 1)

Is this because we're using extend self in there?

mentioned in issue #3326

It was. I expect this to pass now, and have created an issue to address the remaining problems with LicenseHelper: https://gitlab.com/gitlab-org/gitlab-ee/issues/3326

@DouweM can you review https://gitlab.com/gitlab-org/gitlab-ee/commit/8055fdddca7eb9363bc367ca7dd8a66393f9e470 ?

mentioned in merge request !2807 (merged)

assigned to @DouweM

assigned to @nick.thomas

resolved all discussions

added 1 commit

• c11ba136 - Remove broken memoization in LicenseHelper#show_promotions?

Compare with previous version

assigned to @DouweM

@DouweM OK, change in https://gitlab.com/gitlab-org/gitlab-ee/commit/c11ba1366ba88f5e7f46df1cf6f391045a86e98a

mentioned in merge request !2806 (merged)

approved this merge request

merged

mentioned in commit 0ae313c4

mentioned in commit ff0ed5f0