Not quite off the beta branch yet, but I’m hoping that either tomorrow or the next day will see that. There were some severe regressions and some other major issues that just got missed in the last build, but those are all cleared up now and I’m currently not aware of any major bugs in the single player game for the base game and the first two DLCs.
If there’s more I’m missing, please continue to let me know!
Multiplayer remains offline in the beta branch, but I aim to get that sorted out tomorrow. I had been hoping to get that all done today, but the vast majority of my day got taken up by the other bugs on the roster. For multiplayer, the main thing that I have to get working again is the initial handshake and the way that deserialization buffers work; during the refactor, I changed that back in July, and I recall thinking “oh yeah, I’ll have to check and make sure that all does right when I get to testing multiplayer.” At the time, I thought that was going to be August!
Still, as extra-long as this giant beta period has been, the game is exceptionally polished from it. The extra throughput that we have made a severe memory leak I introduced a couple of days ago something that almost flew under the radar. That’s maybe a strange flex, but in the past, or indeed in most programs, a memory like that would have been paired with far more substantial general program slowdown. (I’ve thoroughly checked this for other leaks, and beyond that one boneheaded one from the 17th, we’re good).
I’ve done a couple of major system refactors in my past work, twice on really large business systems spanning a bunch of servers, and a few times on games for Arcen (including the original migration from my hand-coded engine over to have Keith and I piggyback Unity back in 2010), and this has by far been the largest. The entire AI War 2 project has in general been a time of pushing technical boundaries in order to open up new gameplay and modding and performance abilities, and that has been a challenge to balance along with the regular schedule for the game, but I’m really glad to where the technology has (very nearly) arrived at now.
When we first set out to work on AI War 2, neither Keith nor I had ambitions for nearly such complex factions, AI that would process so much data, or support for mods of such complexity. But that’s how the game and its community have evolved, and the modern version of the game really handles all of it very well.
More to come soon.