On Mon, Mar 24, 2008 at 3:16 AM, Jan Hoevers <[EMAIL PROTECTED]> wrote: > Obviously it's Ermal who decides how to release his features. Although I > would like support new features, it's actually the lack of a wide user > base that keeps me away from special versions. As I see it, that is the > down side of this bounty system. > Rolling my own version wouldn't solve that.
Correct, it wouldn't solve it, you'd be running a non-standard release. I do want to clarify that the bounty system doesn't specifically exclude people, nor does it dictate a timeline for when something is released to the public. Large feature changes like the shaper are things that normally wouldn't see the light of day until we're into the next release cycle as they have large enough impact to the tree that we aren't going to inject them into a release that's already in the beta stages (which is actually why the shaper changes never made it into 1.2). At any rate, I suspect I'm arguing a moot point, I understand your concerns with the bounty system, I just don't think it's as dire a situation as you describe. I'll leave this thread with one final thought. In all the bounties I've participated in, I've rolled custom images for my own testing that I've then given to the sponsors of the bounty (it doesn't take me any more time and it makes them happy, even if they were willing to wait until a snapshot). In some cases, the code was also in the repo and being rolled into pfSense snapshot images. One case in particular, the sponsor couldn't use a snapshot since it wouldn't boot on his system (later fixed, unrelated to the code I was working on), so to get the code tested I had to release custom 1.0 images. I don't see anything wrong with the practice and the end user doesn't have to run the code in production (and certainly shouldn't until they've done their own round of regression testing). --Bill