What version are you running?
Describe the enhancement and the motivation for it.
Perforce is a client-server SCM system. Each file diff generated by
rbt postincurs latency of ~1s, and a single changeset of N files requires N
p4 printcallouts in serial. So, a change with ~25 files burns ~30s generating diffs.
It'd be great if
rbt post, instead of generating diffs serially, perhaps tossed all of the
p4 printcallouts (or the entire "writing X to tmpfile; p4 print; diff tmpfile $output" sequence) into a threadpool (default size: ncpus?) in order to parallelize this task.
Please provide any additional information below.