Support cancelation of job script with resumed execution of later stages
What does this MR do?
- Cancellation before the main script cancels the whole job.
- Cancellation during the main script cancels just the main script, with after_script and subsequent stages (such as artifact uploading) working for a
failed
script.
Why was this MR needed?
We want to support two modes of stopping a job: cancel and abort. Cancel should still allow artifacts to be uploaded, and for after_script to always run.
What's the best way to test this MR?
What are the relevant issue numbers?
Closes #30827 (closed)