Fix sporadic failures in t/01_run.t

Wait until all sub processes have been forked before stopping the main sub
process. This may help to avoid sporatic failure we sometimes see on OBS on
ppc64le and aarch64 builds (probably because those machines tend to be
slow).

Related ticket: https://progress.opensuse.org/issues/178183

Author: Martchus

t/01_run.t

@@ -467,6 +467,10 @@ subtest 'process code()' => sub {
   is $p->read_all, '', 'stdout is empty';
 };
 
+sub _number_of_process_in_group {
+  scalar(split "\n", qx{pgrep -g @_} or die "Unable to run pgrep: $!");
+}
+
 subtest stop_whole_process_group_gracefully => sub {
   my $test_script = check_bin("$FindBin::Bin/data/simple_fork.pl");
 
@@ -497,9 +501,13 @@ subtest stop_whole_process_group_gracefully => sub {
   #       stop would also stop the test itself.
   my $test_gpid       = getpgrp(0);
   my $sub_process_pid = $sub_process->pid;
-  sleep 0.1 while $test_gpid == getpgrp($sub_process_pid);
+  my $sub_process_gid;
+  note 'waiting until process group has been created';
+  sleep 0.01 while $test_gpid == ($sub_process_gid = getpgrp($sub_process_pid));
   note "test pid: $$, gpid: $test_gpid";
-  note "sub process pid: $sub_process_pid, gpid: " . getpgrp($sub_process_pid);
+  note "sub process pid: $sub_process_pid, gpid: $sub_process_gid";
+  note 'waiting until all sub processes have been forked';
+  sleep 0.01 until _number_of_process_in_group($sub_process_gid) == 3;
 
   $sub_process->stop();
   is $sub_process->is_running, 0, 'process is shut down via kill_whole_group';