Hi Florian, * Florian Kriener <[EMAIL PROTECTED]> [2008-07-03 00:44]: > On Tuesday 01 July 2008 22:45:09 Nico Golde wrote: > > Why don't you reorder your runlevels symlinks then? > > Because it seems not very sensible to me to start the mta and spamc before > the > network gets started.
That was not my suggestion. > > > So I investigated > > > that problem and found try-restart behaving like it does, > > > starting the daemon if it is not running already (which, > > > in my opinion, is not intuitive). > > > > it's restarting the daemon or starting it if it's not yet > > running. > > Yes, but it's called try-restart and not try-restart-or-start. But this is a > mere sidenote. You are right, the start mechanism should not be there, also in my opinion. It turned out that this was a fix by Hector for #268346. Looking into the bug report it seems like the fix was wrong and the resolvconf script should just call start instead of adding start to try-restart. > > I am not too happy about including this patch for this > > corner-case. The init scripts in Debian include an LSB > > header that defines dependencies on other init scripts (well > > facilities to be precise). Did you already have a look at > > the insserv package which automatically reorders your init > > scripts based on these dependencies. > > > > This might be the right solution for you. > > Please see above, for why that is not the right solution. And I can > understand > your point in not wanting to include that more or less crude patch. But let's > have a look at /etc/resolvconf/update-libc.d/fetchmail: > > if [ -x /etc/init.d/fetchmail ]; then > /etc/init.d/fetchmail try-restart > fi > > I don't think that resolvconf should start any daemon, but it does. So > starting my network causes the start of fetchmail. (BTW: init tries to start > fetchmail too, which then fails of cause). Well these files are treated like any other configuration files as far as I know, feel free to modify or remove them. Hector, what's your opinion towards the fix of #268346 and this bug? Kind regards Nico -- Nico Golde - http://www.ngolde.de - [EMAIL PROTECTED] - GPG: 0x73647CFF For security reasons, all text in this mail is double-rot13 encrypted.
pgpP7GiPJvdt5.pgp
Description: PGP signature
