It turns out that this way of doing things is not really feasible, given the design of rmw_wait. In particular, because this is a per-context condition_variable, it can spuriously wake us up for things that we are not interested in. That would require us to do a check to see if we should go back to sleep, waiting on something we do care about.
Because of that, I'm going to close out this PR. #205 is the start of a series of PRs that should fix this in a different way.
It turns out that this way of doing things is not really feasible, given the design of
rmw_wait. In particular, because this is a per-context condition_variable, it can spuriously wake us up for things that we are not interested in. That would require us to do a check to see if we should go back to sleep, waiting on something we do care about.Because of that, I'm going to close out this PR. #205 is the start of a series of PRs that should fix this in a different way.