Scan results come with a moderate degree of natural noise (timeouts, load-balancing etc.). There should be a simple way to refine a regression result by re-scanning the regressing hosts from a previous regression run with more relaxed parameters (longer timeouts, more retries, less parallelism, etc).
This could be implemented as:
Adding an option to srcupdate to turn the hosts in an existing regression log into a sources database which then can be used for subsequent regression runs.
Adding an abstraction to the -s --source parameter that accepts log handle references, so the host database for the run can be extracted on the fly from a log.
Scan results come with a moderate degree of natural noise (timeouts, load-balancing etc.). There should be a simple way to refine a regression result by re-scanning the regressing hosts from a previous regression run with more relaxed parameters (longer timeouts, more retries, less parallelism, etc).
This could be implemented as:
srcupdate
to turn the hosts in an existing regression log into a sources database which then can be used for subsequent regression runs.-s --source
parameter that accepts log handle references, so the host database for the run can be extracted on the fly from a log.