D19514: Add support for detecting avahi crash - cleanup invalid dbus objects in that case

[Stefan Brüns]

bruns added a comment.


  In D19514#424501 <https://phabricator.kde.org/D19514#424501>, @sitter wrote:
  
  > I haven't had a proper look, so only style complaints for now... but...
  >
  > Isn't the problem rather that the daemon shouldn't be randomly restarted on a live system? I.e. this ultimately is an integration issue on the distro-level.
  >  I doubt many if any avahi clients handle this properly and disadvantageous side effects are rather expected. Same as pulseaudio for example.
  
  
  As Avahi is mostly stateless, I see no reason why it shouldn't be restarted after e.g. an update.
  
  > So I am not sure if this is worth handling at all, but if we handle it we should do it properly and republish/rediscover services. The client shouldn't get broken frontend objects just because the backend of the backend restarted.
  
  +1 - especially, as the client side is almost guaranteed to be implemented.

REPOSITORY
  R272 KDNSSD

REVISION DETAIL
  https://phabricator.kde.org/D19514

To: jpalecek, sitter
Cc: bruns, kde-frameworks-devel, michaelh, ngraham
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20190304/0b0fed3e/attachment.html>