Re: thoughts on new modes (Ie)

[Christophe Kalt <kalt@xxxxxxxxxxx>]

Eh, I was stunned by the wave of "yay!", finally someone
against it ;-)

On Nov 25, Ville wrote:
| I disagree strongly. I don't think we want to change anything like this
| after n+1 pre/beta-releases of the 2.10-series already being run.

Why? it's no big deal at all.

| That not being the main cause for disagreeing, the new klud^H^H^H^Hpatch 
| would IMO be too complicated and likely to cause desynchs if not planned
| properly.

there's no way this would create desynchs as it wouldn't be
silent. (e.g. MODE changes will do it).

|           Maybe it doesn't burn that much CPU, but why burn any at all,
| if we don't need to.

Only the last suggestion has to be implemented carefully,
none of the others really have a cost.

| Nor would much be gained with this patch, as a way around would probably
| be found in a matter of days. (On the other hand, I am definitely not
| saying we should stick to a DFIIWBA[*] policy.)

We're already there. (Thinking about such things because we
introduced new features).

BTW, we could go for a partial change (for instance only
add the restrictions when adding a mask)

| It might still be useful to remember that the users are not likely to RTFS
| (or even RTFM) 2.10 specific stuff. They'll learn 2.10 by just what they
| are told. The majority would probably end up banging their heads to the
| well after trying to figure out the mysterious logic behing the connection
| of +e/I and +b/i.

This argument is not relevant.  Users will be assimi^H^H^H^H^H^Heducated.
It's no different than us banging our heads a few years back
to understand why some bans are rejected by the server.

Which brings an interesting point: these are similar changes
designed to save some bw and memory.  (except my suggestions
go a little further and are a little more proactive)