On Sat, Sep 03, 2005 at 06:42:10PM -0300, Henrique de Moraes Holschuh wrote:
> > 
> > Please provide a note in README.Debian that this is to be the case.
> 
> With all due respect, I do not believe that is the kind of information I
> should have on any README.Debian files.
> 
> It really belongs (and is in) the various Debian administration manuals out
> there in the net.

Invoking the timidity daemon manually is a use case that is specific to
the timidity daemon and very few other services.  There is no reason to
have timidity soaking up memory when it is only used on occasion when
there is a demand for it.  That is why I believe it warrants a special
notice, because the current package is counterintuitive in this case.
I'm not stupid, but it took me more than five minutes to find the
solution, and most users would have given up by that point.

> I will see if I can find a way to make it always active if alsa is active as
> well (and people who don't want it can disable the service through the
> standard ways).

Yes, that is reasonable.

> You should never use update-rc.d directly. It is for package scripts.

Excuse me?  Where is that documented?

> Remove all links but the ones in rc0 and rc6 for example, and the package
> will NOT reinstall any links.

I fail to see the distinction between doing this task by hand and
allowing update-rc.d to do it instead.  Does update-rc.d store a state
somewhere besides /etc/rc*.d?  That would be news to me.

Or is it only this issue:
http://lists.debian.org/debian-devel/2000/07/msg00265.html
Perhaps the problem is that update-rc.d removes both the S and K links,
as opposed to only the S links.  Then later, if both the S and K links
are gone, it reinstalls the links, as opposed to leaving it alone if the
K link is still there.  In that case, the advice "remove all links" is
faulty as it will produce the same results as if update-rc.d were used
to remove the service from those runlevels.

Of course, removing the S links but not the K links means a system which
shuts down more slowly...

-- 
Ryan Underwood, <[EMAIL PROTECTED]>


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to