Automatically reuse stashes in declarative pipelines

With https://issues.jenkins-ci.org/browse/JENKINS-45455 in place and that it can retry pipelines from particular stages is great and saves a lot of time.

What would be even greater would be the possibility to configure a declarative pipeline to automatically reuse previous stashes if some file has/hasn't changed.

The way I imagine the feature to work is that you can give data for a `stash` call that it will use to either reuse a stash from the last successful run or compute it again.

Not exactly sure how the syntax for this would look, but would need some way of declaring in the beginning of a stage what variable to look at before reusing a stash, and if needed, jumps ahead to a `unstash` call instead of stashing again.