Kimaya Bedarkar authored 11 months ago and Björn Brandenburg committed 11 months ago

Currently, the file `prosa.analysis.facts.busy_interval.pi` makes the
assumption that the schedule is work-conserving (in the classic
sense). Results from this file are used to bound the service
inversion. Therefore, the service inversion file also relies on the
classic work-conservation hypothesis, which is too limiting in some
contexts.

This patch removes the work-conservation hypothesis from some of the
results in the PI file. Removing the work conservation hypothesis
crucially requires strengthening two lemmas regarding preemption.

We would like to keep the Prosa compilation warnings-free. Complain
when new warnings are introduced.