[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: hard coded locations for monitrc
From: |
Klaus Heinz |
Subject: |
Re: hard coded locations for monitrc |
Date: |
Wed, 18 Apr 2007 23:44:19 +0200 |
Martin Pala wrote:
> However - because of backward compatibility we should probably keep the
> /usr/local/etc in the search path as well, otherwise it may break some
> existing monit installations which rely on it (although the sysconfigdir
> may have higher priority). The search path should be checked for
Up to now, specifying --sysconfdir had no effect at all because not a single
file contained "@sysconfdir@", so I doubt anyone could have used it
successfully.
Anyone building monit from source and using --sysconfdir probably know
what they are doing: no compatibility issue.
As far as I can see, the only case where backwards compatibility could
be an issue would be for someone using a binary package from some
distributor who either used --sysconfdir=/some/directory in the past
(with no effect) or will do so in the future. Thus, with the proposed
changes a new binary package of monit from this distributor would
require /some/directory/monitrc instead of /usr/local/etc/monitrc.
Is this really a case the monit team should worry about? IMHO, this lies
entirely in the domain of the distributor.
> may have higher priority). The search path should be checked for
> redundancy (i.e ignore the sysconfigdir in the case that it is equal to
> some hardcoded path).
As it is now, there are four tests for ~/.monitrc, /etc/monitrc,
/usr/local/etc/monitrc (to be replaced by SYSCONFDIR/monitrc) and
./monitrc.
The only redundancy I can think of would be a specification of
--sysconfdir=/etc. Would it really buy you much to add
if (strcmp("/etc", SYSCONFDIR)) {
... check for SYSCONFDIR/monitrc ...
}
?
ciao
Klaus
Re: hard coded locations for monitrc, Klaus Heinz, 2007/04/17