On August 18th, 2015, 7:35 nachm. CEST, Stephen Gallagher wrote:

src/rolekit/async.py (Diff revision 1)
def error_handler_with_conversion(e):
170	# We can’t use log.exception() because the traceback is no longer available.	170	# We can’t use log.exception() because the traceback is no longer available.
171	# So the three cases in dbus_handle_exceptions amount to just this.	171	# So the three cases in dbus_handle_exceptions amount to just this.
172	if not isinstance(e, DBusException):	172	if not isinstance(e, DBusException):
173	log.error("{0}: {1}".format(type(e), str(e)))	173	log.error("{0}: {1}".format(type(e), str(e)))
		174	if not isinstance(e, RolekitError):
		175	log.exception()

This change really needs a comment to explain it. Why is it okay to call log.exception() if we have something other than a RolekitError here? That disagrees with the comment on lines 170-171.

I can't say much about the previous comment because I didn't write it :). All I know is that I had some issue throwing exceptions on the daemon side and no way to find out where exactly it bombed. Adding the call to log.exception() logged the complete traceback just as I would expect it. Do you have some background for me? Then I could maybe change the existing comment to reflect what's really going on.

- Nils

On August 12th, 2015, 5:43 nachm. CEST, Nils Philippsen wrote:

async: log tracebacks with (most) exceptions