[CF-Devel] combining multipart images

[Mark Wedel]

  I've done some investigation into what it will take to combine the multipart 
images into one part for client display.  What this basically means that if you 
have a 4x4 object (say a shop), you just have 1 64x64 image, instead of 4 32x32 
images.  The main thing this gains is that image maintenance is easier (no need 
to take the step to re-combine and then re-split to touch up an image).  I know 
some people have expressed the obnoxiousness of having to do that.

  Allowing for these combined images actually isn't that much work to do - 
really nothing at all in the server, and some code change necessary for the unix 
clients at least.

  The biggest problem are people using old clients (whether they are the unix 
ones or anything else) - whether they will display the data correctly, or even 
work without crashing (I think some of the very old unix ones don't expect 
anything larger than a 32x32 image).

  My thought on this is to make the new unix client available, and have the 
server send a standard drawinfo message to the user (client) if it thinks the 
client is out of date - basically telling the user to update.  Then after some 
amount of time has passed (few months), people can start putting combined images 
into the cvs repository (for the most part, I don't expect there to be much 
updates, as unless the image is getting worked on, there really isn't much 
reason to combine the images that are already split).  Players that are using 
the old clients at that time would then face the consequences.

  The only way to actually remain backward compatibilty would be to keep a set 
of the old split images around, but that seems to sort of defeat the purpose of 
trying to make things easier by not having to deal with split images (I'm 
thinking of the scenariou of adding a new archetype/image - if you have to go 
out and split up the image, that would seem to make it a bit of a bother).