Re: Re: 4.0.6

At 04:15 PM 5/2/2001 +0300, Zeev Suraski wrote:

At 16:02 2/5/2001, Andi Gutmans wrote: I disagree. We are not getting enough testing of our RCs.

I think if we announce an RC and we tell people they are just helping us test the pre-release it's OK. It's not as if they can't grab a snapshot.

People usually tend to deal with pre-release or release candidate as if they're releases. It's not that RC's are necessary significantly less stable than releases, but I don't think that announcing them on such wide forums would give our QA efforts anything. It's pretty much the same as saying we announce 4.0.{x} as an RC for 4.0.{x+1}, because 4.0.{x} bugs will start flowing as soon as it's out. What I'm trying to say is that if we make that jump from a QA team to the entire world, then essentially, we go a step backwards. I think that the way things are today is good, and most of the bugs which aren't found can only be found in wide scale testing, but I don't think that announcing RCs in prominent places is the way to go. Perhaps we should announce the QA team again, so that people who would join but don't know about it would get a chance.

I don't think that the PHP mailing list is the whole world. It's a fairly large amount of people but a relatively small percent of the amount of downloaders (<10% IMO). I think that the amount of QA people (even if the amount doubles) is enough to find those little bugs that crept in. And many bugs can only be found by trying to actually run the RC for a few days on a development machine. I don't know how many QA guys actually do that. Andi