Add --no-flash option to probe-run

[therealprof]

Skipping the actual flashing process is useful if the binary to be run
is already in flash, either due being a re-run or use of a different
approach to do the flashing.

Signed-off-by: Daniel Egger [email protected]

[japaric]

I think that using --no-flash with --defmt would be error prone. There's currently no way to check whether the passed ELF and what was last flashed on the device match and if they don't match then all defmt messages would be detected as corrupted. So for now (*) I would like a CLI error to be thrown if one tries to use both flags.

(*) at least until we implement a mechanism to detect that what's flashed on the device matches the ELF we are going to use to decode the defmt data

[therealprof]

I don't think it would be a lot more error prone than using without --defmt but I agree that it would be a lot better if we could get some feedback from cargo passed to the runner whether the implict cargo build modified the target or not. I looked around but couldn't find anything in the passed on environment about that.

I can make the two options mutual exclusive for now if you still prefer that.

[japaric]

Thanks!