The test hang appears to be caused because the tests fake the dbus server responses from pid 1. However, with the new API, we would also need to either:
1) Fake the call to job_reload().
2) Call job_reload() explicitly.
3) Change the test to actually fire up an upstart instance to avoid faking anything.
Option (3) is preferred of course, but since test_job.c:test_reload() is already doing (3), it should be sufficient to simply test the client is making the correct type of request over D-Bus for test_initctl.c.
The test hang appears to be caused because the tests fake the dbus server responses from pid 1. However, with the new API, we would also need to either:
1) Fake the call to job_reload().
2) Call job_reload() explicitly.
3) Change the test to actually fire up an upstart instance to avoid faking anything.
Option (3) is preferred of course, but since test_job. c:test_ reload( ) is already doing (3), it should be sufficient to simply test the client is making the correct type of request over D-Bus for test_initctl.c.