Document crate discovery and (arbitrary) experiment creation / execution

[ryankurte]

Hi all,

I'm hoping to use crater to regression test pre-releases of core libraries, specifically at the moment the embedded-hal crate which underpins the entire embedded ecosystem (and thus causes us some release-time anxiety).

The documentation suggests arbitrary experiments can be run over subsets of discovered crates, but, I can't seem to find any further information about how to achieve that (and it is non-obvious from the CLI help).

I think the process I'm looking for would be to pull all crates that depend on embedded-hal, execute a build and test cycle to create a baseline, then patch to a pre-release version and re-run the tests to observe any difference in failures.

I figured before digging into the code too far it was worth asking:

• is this something crater can / should be able to do?
• is there existing documentation / are there examples of existing uses of crater for this?
• is this something that could be documented / could examples be provided?
• are there features we would need to implement to support this use?

Cheers,

Ryan