On 02/10/2009 11:06 AM, Peter Palfrader wrote:
On Tue, 10 Feb 2009, Michal Suchanek wrote:
If you write the proper pid into the pidfile it should all work out. Or
you kill the tor process and start it again.
I think we can close this bug?
Perhaps tor should check the return value of the write()/close()?
Maybe. And then what? Just die? We can't say anything on stdout/err,
we already detached. This isn't as trivial as it may seem.
Log on syslog or just die. Whatever. But if it fails to create the pid
file it is a failure and daemons normally refuse to start in that case.
It can be even done so that the daemon forks and the parent writes the
pid file and if it fails it kills the daemon again.
Note that logging on syslog would work, it runs as root. And if even
syslog does not work you should notice.
Thanks
Michal
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
