Tim Rightnour wrote: > So.. recently.. while writing the attack messages, it was proposed to me that I > should write the messages on the client end and have the server call them with > a smaller bandwidth message. I didn't even consider this because I'm simply > not going to hack N clients to do it. (there are other fundamental problems > with it too.. but I'm ignoring those for the purposes of your query) For that example, my opinion would be it would be in the common client code, so only that would need to get updated - the result here is that it would mean that that one update would result in all clients based off it (gtk/gnome/x11) to work. IMO, these types of changes are not a big problem. > 1) The Xlib client is somewhat unmaintained because virtually nobody knows > Xlib/XT/XAW anymore.. and those of us that do, know better than to admit it. I'll admint I know X11. But it really ends up being such a pain to add most new features with the functions that base X11 offers. The result here is that at least up to current time, the X11 client would work on the servers, but many of the new features offered are not supported (I think it only supports the standard 11x11 map size, does not show resistances, etc). At some point, this becomes a problem, because it becomes desirable to remove the old code from the server, which may then mean that the x11 client really does need to support some of these features. > > 2) The gtk client has a hideous memory leak: > root 24366 0.3 4.9 13860 25840 ?? Ss 26Sep01 96:53.29 /usr/X11R6/bin/X > (that gets to around 70 megs before the whole X server becomes useless) You can try clicking on 'clear info window' under the client subwindow. There may very well be other memory leaks, but the scrollback is the big one that is known but no convenient fix due to bugs in gtk. > IMHO.. All I require is that it runs reasonably fast, and works on 8bit and > 24bit displays. I don't have modern graphics hardware really.. and not > everyone does. Basically.. I just want to be able to play the game, on my > systems. If I can't run the client on NetBSD because of some exotic GL crap, > then I'm going to be upset. (just like if the Xlib map editor goes away I'm > going to be displeased) I think the Xlib map editor will go away - it is already in the mostly unsupported category - the map attributes are no longer modifiable from within the editor. I have a feeling as more things are changed, the Xt editor will become more and more out of date.