You're right... we've even changed the location they are expected in... so we do indeed need our data not just on sourceforge, but properly under revision control...
On the theme, it would be nice to have a 'gae_techs' too, and have an xml only magitech with patrol/gaurd/etc commands...
Yea, as a matter of fact, I believe that I had already begun modifying the magitech faction in the tree on codemonger.org (I'll have to re-check it) just for that purpose. In fact, it was then that I realized that the "auto-return" functionality was a bit of a cheat because it bypassed the "Guard" command that is only supposed to be available after you have researched "training field" or some such. The main aim when I made the auto-return functionality was really to not have them automatically run to the location where they 1st saw an enemy and auto-attacked it, this was very messy. The correct behavior is for them to stop in their tracks after having auto-attacked, unless they are executing some guard or patrol command.
Anyway, that's OT, sorry.
There were also changes to the language files and something else, I forget what
What if the data is kept up-to-date with trunk and then every time a tag is created you copy the data into the tag? It still leaves manual modification of branch data changes though. But how often will branches be breaking compatibility?
Would a setup like this be better?
Actually, I think this is something that needs to be done completely accurately. There is a very specific purpose for trunk, branches and tags that subversion defines. It might be a good idea to read up on
Subversion Best Practices. Also, a lot of the bandwidth problems can be mitigated by using "svn switch" to switch your working set between branches. This will only download the files that are different and all of the big data files that haven't changed will not get downloaded again.
