At 10:35 AM 5/25/01 -0400, you wrote:
>* Make JUPE, GLINE, OPMODE, and CLEARMODE show oper nicknames when
> HEAD_IN_SAND_SERVER_NOTICES is #define'd
This would seem to be me to be better as an option for the F:'s.
Default to showing the nicks when server notices are restricted, and
hiding them when they are open, but allow an F: to override.
>* Privilege-ize HEAD_IN_SAND_*
>* Remove old chroot code
>* Remove JUPE/GLINE rebroadcast code--preserve BURST behavior, though
>* Add a flag to SERVER to mark hubs as such
Should be configurable, if we want to continue our philosophy of
helping the smaller nets. Many small nets with less resources than
Undernet have to have servers do both client-server and hub duty.
In this case, if these networks also leave MAP and LINKS available,
it could lead to misleading situations. In any case, it looks like
it would be quite a simple thing to have a configuration option that
tells whether or not a hub will actually report itself as such.
Again, seems like something for the F:'s, such as having both, say,
HUB and REPORT_HUB_STATUS perhaps? ...
Also, something I suggested on #coder-com, is a flag in SERVER to
report a sum code based on the HEAD_IN_SAND_* options defined. As
these are meant to be network-wide policies, they should be the same
across all servers. Therefore, it is reasonable to have a coded-in
number, unique for any particular combination of H_I_S options, that
is sent with SERVER. Then, if a server sees a connect attempt from
a server with a different policy code from its own, it can reject
the connection, indicating that ircd_policy.h needs to be checked.
This makes the policy system tight, since obviously a single server
that chooses not to use these #defines leaks all the information
unet-admins@ wants hidden.
Since each HEAD_IN_SAND define is boolean, the simplest way to do
this would appear to be to have a single-bit value associated with
each such define, then create the code from a simple addition of
these values.
