URL: https://github.com/SSSD/sssd/pull/94
Title: #94: Enable {socket,dbus}-activation for responders

pbrezina commented:
"""
Squashing is fine, of course. BTW I'm not quite sure if this version contains 
the uid/gid changes as I see it as parameters in all systemd unit files.

In my opinion, even private communication between responders and data providers 
should prolong the timeout. The communication is triggered by the client 
anyway. The basic idea is that when a client contacts responder, the responder 
contacts data provider and awaits reply. It may take a long time for the dp 
reply to come, but we are still not idle, we await a reply. It probably won't 
make much difference since we send reply to client as fast as we can when we 
get reply from dp so the timestamps will be similar, but we don't have to 
special case IFP anymore and I like that :-) What do other @SSSD/developers 
think?

The only incoming signal SSSD can handle right now is NameOwnerChange and this 
should not take part in the decision whether SSSD is idle or not. All signals 
are handled inside `sbus_signal_handler` and should not make it into message 
handler, if the do I'd consider it a bug. 
"""

See the full comment at 
https://github.com/SSSD/sssd/pull/94#issuecomment-264175621
_______________________________________________
sssd-devel mailing list -- sssd-devel@lists.fedorahosted.org
To unsubscribe send an email to sssd-devel-le...@lists.fedorahosted.org

Reply via email to