Control: tag -1 upstream patch
Control: forwarded -1 https://github.com/martinpitt/python-dbusmock/pull/245

Hello Simon,

Simon McVittie [2025-07-14 14:02 +0100]:
> After my recent upload of mutter 48.4-2 to unstable, its autopkgtest 
> failed on ppc64el. At least one of the failures (monitor-unit.test) had 
> this logged, and the test subsequently timed out:
> [..]
> > dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoServer: Failed 
> > to connect to socket /tmp/dbusmock_data_nh20oy7q/system_bus.socket: 
> > Connection refused
> 
> I think this indicates that dbusmock started a dbus-daemon to be the 
> mock system bus, but then attempted to connect to the socket before the 
> dbus-daemon was listening on that socket. There is a (very) short window 
> of time between bind() and listen() during which connection attempts 
> will fail.

You are spot on. There were *two* implementations of dbus-daemon startup in
dbusmock's code, and the good one became dead code a while ago, and the bad one
was the live one. Fixed and cleaned up in
https://github.com/martinpitt/python-dbusmock/pull/245

Thanks a lot for the report and analysis!

Pitti

Reply via email to