ctld: Move the pidfile handle out to a global variable

[jhb]

Summary

This ensures it will be destroyed (removing the associated pidfile)
anytime the process exits, including from exit(3) calls. This fixes
a few places that would "leak" the pidfile on certain errors.

This also removes the need for some convoluted logic where
configuration objects would hand-off ownership of the pidfile handle
from the old configuration to the new configuration.

Sponsored by: Chelsio Communications

[asomers]

I guess this is ok, even though I normally try to avoid globals if possible. pidfile handling could also be simplified if we forbade changing the pidfile's location during a config reload. Do you think that would be acceptable to do?

[jhb]

In D56527#1293831, @asomers wrote:

I guess this is ok, even though I normally try to avoid globals if possible. pidfile handling could also be simplified if we forbade changing the pidfile's location during a config reload. Do you think that would be acceptable to do?

I have some later commits (just posted) to make some of the logic around pidfile renaming less grotty. That said, it is not at all clear to me that changing the pidfile's location is useful. I would personally be fine with just ignoring pidfile pathname changes or rejecting config files that try to use a different pidfile path than the current path.

I'm not sure we can easily get away from the global. Maybe it could be a local variable to main(), but handle_connection() still needs access to it somehow. You could make it a static member of the conf class maybe but that's really just a global with a different spelling.