[Bug 135393] Re: init.d-script uses control channel instead of signals

2007-10-03 Thread LaMont Jones
** Changed in: bind9 (Ubuntu) Assignee: (unassigned) = LaMont Jones (lamont) Status: Triaged = Won't Fix -- init.d-script uses control channel instead of signals https://bugs.launchpad.net/bugs/135393 You received this bug notification because you are a member of Ubuntu Bugs, which

[Bug 135393] Re: init.d-script uses control channel instead of signals

2007-10-02 Thread LaMont Jones
Generally speaking, signals are a very poor choice for communication with an application for anything other than 'Die': No actual processing can be done in the signal handler, since signals are delivered asynchronously, and even people who know how to safely use signals are frequently reminded

[Bug 135393] Re: init.d-script uses control channel instead of signals

2007-09-10 Thread LaMont Jones
The control channel was added specifically to allow more commands, and to avoid the race conditions that are so easy to create when writing code that uses signals. It's far better to restrict access to the control channel, than to use signals. I doubt that I'll ever change the init script back

[Bug 135393] Re: init.d-script uses control channel instead of signals

2007-09-10 Thread LaMont Jones
The control channel was added specifically to allow more commands, and to avoid the race conditions that are so easy to create when writing code that uses signals. It's far better to restrict access to the control channel, than to use signals. I doubt that I'll ever change the init script back

[Bug 135393] Re: init.d-script uses control channel instead of signals

2007-09-10 Thread Toni Heinonen
I'm sorry, but what more commands do you require besides reload, restart, stop, start and status that the init.d scripts give you? And the race conditions come from threads, or? named is just one process. -- init.d-script uses control channel instead of signals