We used Sourceforge's Bugs and Tasks(when sourceforge worked well) and people just stopped using them. We had BugZilla, no one used it. We have a -private- cvs, that we cannot give access to everyone. That is why we set up a bidaily export to a public CVS for everyone. The problem is we used sourceforge's CVS and it's screwed up. We do NOT however need people running CVS code if they do not know what they are doing. You see the guides that give CLEAR STEP BY STEP instructions for compling from cvs? Yeah well there are tons of questions about people that cannot even get cvs setup to begin with. Those are not generally the kind of people that will a) report bugs in a way that is USEFUL (omg it not work fix it plz!!!111). or b) able to fix the bugs themselves. If you would even read trumpcard's post he said the same thing.
I will explain our development cycle, and other issues as best I can.
1) Start, mostly working code that is current with live
2) Fix bugs or add features to this code -OR- Have to update everything to live
3) Internal testing
4) Release Development Releases(DRs) of the next version, typically have at the very minium walking around and zoning, but normally have 80% of previous feature set working and everything important working.
5) Continue updating to live or adding features AND fix bugs that are reported in a USEFUL way.
6) Spend plenty of time reading posts or trying to fix someone's issues and it turns out they screwed up and are using a bad database, or doing something stupid.
7) Reply to posts about people complaining about AAs not working right or pathing, state that we need people to packetcollect AAs and send us the logs and never receive any logs.
If ready with what we believe is a STABLE version that works with live and has 95% functionality, release new version
9) Goto 1