Signed-off-by: Stefan Büringer buringerst@vmware.com

…2944-to-release-0.19

[release-0.19] 🌱 Verify PR titles with shell script

This updates the fake client to retain the PartialObjectMeta TypeMeta
when getting resources.

…2949-to-release-0.19

[release-0.19] 🐛  Fakeclient: preserve TypeMeta for PartialObjectMeta resources

Co-Authored-By: Aviv Keller <redyetidev@gmail.com>

…2961-to-release-0.19

[release-0.19] 🌱 pr-verify: use env var for passing the PR title

Co-Authored-By: Aviv Keller <redyetidev@gmail.com>

…2964-to-release-0.19

[release-0.19] 🌱 pr-verify: use env var for passing the PR title

Signed-off-by: Stefan Büringer buringerst@vmware.com

…2973-to-release-0.19

[release-0.19] 🌱 Fix PR verify action

The fake client currently has a number of time of check time of use
races, where it fetches an object to determine what to do in a mutating
operation. The problem is that the object might change in between
fetching it and doing the mutating operation. Most notably, this
happens when:
* Patching is done in parallel. Only one of the patches will succeed,
  the other ones will fail with a conflict
* Updates of objects that allow unconditional updates: All updates will
  succeed, but not all of them will increment the resource version (i.E
  dirty writes for the RV)
* An update for an object that allows createOnUpdate races with a create
  or delete
* A DeleteAllOf call races with Delete calls
* A scale update races with a normal update

This change:
* Adds tests for all of these cases
* Fixes them by adding a lock around the write operations, including
  their read part, if any

…2980-to-release-0.19

[release-0.19] 🐛 Fakeclient: Fix TOCTOU races

…3018)

* Ensure all WatchFunc enable watch and boomarks

AllowWatchBookmarks is generally pretty safe to enable as it has been
available in Kuberentes for a long while, and the server ignores the
flag if it doesn't implement it (per docs).

Signed-off-by: Vince Prignano <vince@prigna.com>

* Defaults to false for 0.19

Signed-off-by: Vince Prignano <vince@prigna.com>

---------

Signed-off-by: Vince Prignano <vince@prigna.com>
Co-authored-by: Vince Prignano <vince@prigna.com>

This change makes the leader elector use a client that internally has
a smaller timeout than the renew deadline, which avoids a situation
where a single request timing out makes us lose the leader lease.

Signed-off-by: Vince Prignano <vince@prigna.com>

This change is a follow-up to the one that introduces the usage of the
leader-elector with client timeout. That change was breaking because it
introduces a new option and always assumed it was set. This change makes
us only use that option if its actually set.

… of fsnotify (#3023)

* Reestablish watch for the certificate paths

* Remove fsnotify and use cached read watcher

* Simplify return

* 🐛Fix certwatcher test to be backwards compatible

---------

Co-authored-by: Maxim Muzafarov <m.muzafarov@gmail.com>

…3028-to-release-0.19

[release-0.19] 🐛 Use leader elector with client timeout

Signed-off-by: Vince Prignano <vince@prigna.com>

🌱 add deprecated Watch() to certwatcher