[ircd-ratbox] ratbox-services chanserv lax on flags.

Jilles Tjoelker jilles at stack.nl
Tue Apr 22 20:08:47 UTC 2008

On Mon, Apr 21, 2008 at 09:10:52PM +0100, Lee H wrote:
> On Sun, Apr 20, 2008 at 11:39:01PM +0100, Chris Rees wrote:
> > I'm tweaking ircd-ratbox and ratbox-services etc, and find that when I
> > restart both (ircd-ratbox + ratbox-services), chanserv neglects to
> > apply modes, for example +i #operators.

> > This means that after a restart, channels are unmanaged until a
> > registered (channel) op jumps in, and sets the modes, which are then
> > enforced.

> > Why does it not seem to obey CREATEMODES, or have I misunderstood them?

> CREATEMODES is only used if chanserv is supposed to be in the channel
> generally (i.e. AUTOJOIN is on), or it is forced to join the channel for
> some other reason (i.e. to ban/kick or deop the only user in the channel,
> when a ban matches, or RESTRICTOPS/NOOPS are on respectively).

In fact, CREATEMODES is only used if chanserv is creating the channel or
lowering TS, both of which only happen if AUTOJOIN is on.

Note that if ENFORCEMODES is set, it is used and CREATEMODES is ignored;
a case could be made to OR them in that case, allowing for some modes to
be set by default but alterable by any chanop while forcing other modes
on at all times.

> There is a further conf option to chanserv called "autojoin_empty", which
> controls whether chanserv will create channels marked as AUTOJOIN.  The
> default to this is off -- as being enabled tends to result in lots of
> channels that have noone but chanserv in them.

> It would appear that autojoin_empty is missing from the example.conf though.

> If a mode is to be enforced, ENFORCEMODES should be used..

Yes. ENFORCEMODES is analogous to MLOCK in some other services packages,
except that you cannot enforce that certain modes are not set. (You can
however ensure a channel is recreated without +n and/or +t, by enabling
AUTOJOIN and setting CREATEMODES and ENFORCEMODES such that they do not
include those modes.)

Jilles Tjoelker

More information about the ircd-ratbox mailing list