Michiel Meeuwissen wrote:
Rico Jansen wrote:
START OF CALL: 2004-09-22
END OF CALL: 2004-09-27
Accept the new multicast/unicast code:
[X] -1 (NO), because :
No I don't accept the code as is , since by removing the spawn per message
it will break when a message gets handled incorrectly. When an exception
occurs for a local message, future messages both local and remote won't
be handled.
I have not looked into the code, but the logic of that escapes me. If an exception
would be that
fatal (e.g. in some daemon), it should be catched.
It isn't for local changes, it is for remote, so that is easy to fix.
But that doesn't eleviate the problem of slow or hanging changes.
--
Rico Jansen ([EMAIL PROTECTED])
"You call it untidy, I call it LRU ordered" -- Daniel Barlow
