Improving test starting from a bug

[sxd]

Currently, since we implemented the /ready-to-test line which use pull_request_target instead of pull_request to run the test, we're expecting the label ready to test :elephant: and we cannot see this when running the test on pull_request or even when running the /test command.

The proposal is:

• Only run on pull_request_target when label is equal to ready to test :elephant: and not on any other situation, even when labels are removed.
• The /test or /ready-to-test command should be used to run all the E2E test, unit test can stay for now to run on every situation.
• Even if members of the project create a PR this will not run until /test or /ready-to-test command are trigger.

The idea with this approach is to:

• Do not run test every time we create a PR or we push
• Avoid the excessive consumption of runners, since we don't have so much runners
• Run test in PR from forked repo

What do you think about this @phisco @leonardoce @gbartolini ?

[sxd]

Another issue spotted here, every time we rebase a PR on main or we push a rebase, we trigger the pull_request action here, meaning that we ask for a new run and most of the time we don't need a new run, we just one to merge because we already had a test that it's actually working