Yes! Before the audit we had a release manager. It was his responsibility to track down which code segments worked and test them. I guess right, that was Ge job at one time when we had cvs. Things have slimed down since then and many developers left the project. I guess now due to laziness on our part this has been unfilled. Since we have svn now we do branches and that job went to Brandon, he's gone too so~. Yes we are lazy and the easiest fix to all of this mess was to create a release branch and sync fixes to that like I use to do. Basically doing the thing Aleksey wants to do in the first place with out starting a fight.
On Fri, Jun 20, 2008 at 5:51 AM, Richman Reuven richman.reuven@gmail.com wrote:
hi! i've been following the project for a couple of months and this kind of drama, i really didn't expect...
both sides have a point. a serious project needs a working tree and a work-on tree. (the temp working implementation vs correct implementation by magnus isn't a simple issue, perhaps there should be another tree for that as well or for it to be a work-on-far tree like linux-staging tree - to work on things which don't enter the next release...) but asking one developer (and quite an experienced, one of the top contributors) to be the only one to be double checked is... coding conventions also do matter. however I don't think that the diff between !f00 and f00==false is what's causing the breakage.
i really don't understand how such a project at this stage, doesn't adapt any of the methods used by virtually every other project to manage the process... also as far as the "it's a private problem" opinions, it couldn't be any more wrong! this kind of thing and these decisions will be _hunting_ the project for a long while! i know both sides seem to be in a "pinch" but it'd be really best if you guys said "sorry i was wrong" in private and reported the issue as resolved...
sorry for the intrusion...
No,,,, good one~