Make wait*() behave the same as it did before we moved TSTOP handling
out of tsleep(). The behavior was that if both parent and child are
stopped together (aka via ^Z), and then resumed together, and the parent
is doing a WUNTRACED wait*() on the child, the parent will only see the
resume. Without this little hack the parent will see both the STOP (after
the child has resumed since the parent was stopped too!), and then the
CONT.
It is unclear whether this is mandated by standards or not, but for now
we are going to maintain backwards compatibility.
Reported-by: YONETANI Tomokazu <qhwt+dfly@les.ath.cx>