...and now people want 0.76b2
The question usually is whether it is needed or not, and I like to answer it isn't, but I think in this case there is reason to start working on a bugfix release.
To look back on the 0.76b1 release, I think it went very well all in all, even if I say it myself. Really started on the release sometime around 14'ish, installer up around 16:30'ish and all critical stuff was up 'n running before 18:00. I think that's pretty fast compared to earlier releases, and I have the feeling the "buildbot" way of doing it is starting to pay off. (including the testing, etc.)
Two of the major points (of failure/attention?) in the released stuff where not really in the engine itself, but more in total system integration etc.
One is the ground decal issue in settings.exe, which has been solved already, but (unfortunately) is already enough to warrant a b2 release. This should be a good example to everyone to document more: had the setting be documented clearly this probably wouldn't have happened. (and yes I know I'm guilty too )
The other one is the new installer. While it's all nice and dandy, before b2 we do need to set up proper load distributing on the server (should be: serverS) behind it. In terms of server load it worked fine since it was just serving files and not doing CPU intensive PHP stuff or anything, but pipes were saturated and as such download rates were very (unacceptably?) low.
(For the stat whores approx. 660 unique IPs hit the server through the installer just on monday after release (ie. between 17:30(?) and 00:00), and some 400 more on monday night.)
On top of this the datacenter is down now, so the installer is broken. It would be cool if we could prevent this too (redundancy), but I think this may be a bit too much to ask
I should've realised this all before b1 of course, but I failed at doing that
Then, we get to Spring issues.
The number of major bugs that slipped through testing is fairly low in my opinion. We do have some 20 odd new bugs reported on mantis, but I think this is mostly because of the promotion of mantis as bug tracker recently, and not because the build is actually more buggy then with earlier b1 releases. If you think otherwise however, feel free to say so
The open bugs in the engine that would be nice to fix before b2 is, AFAICS:
- #757 (spec timeout causes host crash) (mostly fixed before b1 but apparently it can still happen in certain circumstances)
- #739 (hosted replays have a bug)
- #759 (players joined under wrong names/identities)
- #770 (Pings are high until you change the speed of the game)
As for feature freeze, I don't want to demotivate people by holding back development too much, so I won't force a total feature freeze until a few days before b2 is really going to come. I trust you all to take a bit of responsibility and not commit your uber-I-refactored-all-of-spring-and-probably-introduced-lots-of-bugs patch yet (or test it a little bit more)