Re: PID-reuse races fix, introduced GCC validation brakage

[Yvan Roux]

On 30 March 2016 at 23:06, Ben Elliston <address@hidden> wrote:
> Rather than killing each process by PID, what about killing the whole
> process group?  Wouldn't that be cleaner?  Is it portable enough?

In that example the GPID is the one the make check-gcc command, so if
we kill it the remaining testcases to run will not be run, and killing
the group of processes with the negative pids passed doesn't work.  I
agree it would be cleaner if we can do it, but I didn't managed to
find a better solution (yet) than gathering all the childs pids and
killing them.

> Ben