> I don't think you understand what you're doing.  You are not sending the
> information for short packets, only with the text sent for legacy clients.
>
Your solution for short packets is a good one.  I didn't include the
genspkt.c file, but I would not change anything there (this patch assumes
your patch has already been applied).  Both short and long packets are
successfully sent to client with the new army doosh info.  I have tested
it on the client I work with.

As for the "don't change legacy format" argument, if it doesn't break a
single client, and only makes old clients better (by giving them planet
doosh info), where is the harm?  If we continue to just change short
packets for all new info we send to clients, we force people to upgrade to
the newest clients to get the informational advantage that comes with
server changes.  Maybe we want to do that, force people to upgrade.  If
that's the consensus, I'd accept it (my personal opinion is we should not
force upgrade).  If I'm playing 10 year old COW, or some other really
old client without short packets, and I had the choice between the server
sending me info on planet doosh, vs not sending, why would I ever choose
to not get the doosh info?

> You are also annoying re-indenting the code.  Why is this so hard for
> people to get right?  Changing two lines of code without changing the
> indention of the dozen lines you didn't touch shouldn't be hard.  darcs
> even has its nice interactive patch selector.
>
Indentation looks fine to me.  Everything went in 4 spaces because of the
new if else loop to check if armies == 0 or not.  Similiar to how the code
looks in the player kill section.  Unless it's a problem with tabs vs
spaces?