Re: CVS Account Request

> Secondly, TONS of bugs have been introduced by buggy CVS commits from > people that touched things they weren't completely familiar with, or worse, > did things in apparently non core places, that still resulted in either > security problems or other sorts of crash bugs. I'm not sure where we > would have been without Andi's "CVS police", and frankly, I have no idea > where he's drawing the strength to go over almost all of the CVS commits from. I have not seen anything during the last three years which would qualify as "tons of bugs" introduced by committers. It is also the first time that I hear the term "CVS police". We all do peer review; that is all php-cvs is about. People, no matter how long they have CVS access, have a tendency to do early commits. But frankly, I don't think that is caused by people leaving their areas of expertise. That is a people issue though and cannot be fixed by technical measures. > Now, one can make a big deal out of it, and take it 'personally', that I > don't think the Web guys should have write access to the source code. This > shouldn't be the case though. It's just better for things to be organized > than not. Since Zak, for instance, has no benefit in having write access > to the php4 tree, there's no good reason he should have write access > there. I wouldn't deny him, or anybody else access if it's useful for > them, but right now, people are getting access to the source repository for > no good reason other than a historically messed-up setup. We can regulate the access to various modules using ACLs. No need to divide repositories per se. Having said that, I see currently no reason to regulate that much further. - Sascha