@Karl: Well, if Jani would have just created a new branch and committed the patch, I would still be upset, but I could follow your line if argument a bit more easily. However I still maintain an email could have been send to get a more productive result. But what Jani did was commit something that both technically and organizationally was unacceptable against a direct RM decision (I should add that several other contributors agreed with this decision .. actually I only remember Jani opposing that decision).
Anyway, at this point its clear that nothing will happen to Jani's karma. The only person able to do this sort of decision (especially in regards to such a long time contributor like Jani) would be Rasmus unless there would be universal support among the core developers. But Rasmus prefers to run things differently, though in the past years he has mainly been watching from the side lines and intervening with a few emails here and there. So Karl all in all you have little to worry about the core process becoming too restricted any time soon. However I also think that the ability for the community to get their voices truly heard are limited .. even for those willing to speak with C code exactly because of the lack of processes.
This blog post for has served its purposes: one to just be able to do something about my anger and second to get things documented. So it goes ..
Maybe the time has come for a distributed versioning system?
Lukas, I think you missed my point. I'm saying Jani's actions were a symptom, and by continuing to focus on just his actions, you're ignoring the broader problem to the detriment of the community.
The RMs handed down decisions from on high.Those under the RMs either rolled their eyes, gave up on what they were working on because it was no longer interesting, or in Jani's case, staged a protest and threw a molotov cocktail.
You said yourself that HEAD was a "motivation killer", "does not hold enough merit to the world in general", and made several other statements about stalled development ... while simultaneously admitting that the RMs were stalling development and trying to direct all efforts towards something that people find merit-less. Do you see the problem there, or do I need to spell it out for you?
I don't think a distributed development system or any other technology would solve a problem here. There seems to be an issue with the leadership of the internals team. I respect everyone on that team greatly, but from outside the team and in the trenches and foxholes with the rest of the user community, I say: QUIT IT, ALREADY!!! Jani is acting out, yes, but it's a symptom of a larger problem. Get past Jani. Find that larger problem and fix it so that we can get on with things and so that you don't lose the entire developer community to terminal boredom and have PHP6 become synonymous with Duke Nukem Forever.
@Karl: I am with you on a lot of points. Except I do not remember where I said that the RMs were stalling. If you are referring to the decision to not merge the output buffering patch, then this was really not so much an RM talking down to the lowly developers issue. Most core developers agree, like I said in my post, it was really only Jani that repeatedly lobbied for its inclusion.
We have all sorts of organizational problems in PHP, I am the first one to admit that and as I also stated I was expressing my sympathy to Jani because PHP6 was stalled. But its not the fault of any RM that this was the case and it didn't take a revolution to "un-stall" it. We decided to give current HEAD another shot last summer, nothing happened and so the decision to drop current HEAD/trunk was looming in the air. Like I said I was actually trying to work out an RFC for this a few days before this happened. Of course Jani didn't know about it and he isn't the type of guy that sits down to write an RFC, but acting out twice in a row on a single day isn't legitimate .. especially because he is one of the old dev's. I think he has plenty of other ways at his disposal to make himself heard and bring about change but he chose to use an unacceptable option. And this is what this blog post was addressing mainly. The other part was hoping to convince some people that adding some rules could potentially empower new people so that they can more easily become part of the development process.
Sure today php.net has a github mirror but it's basically only used as a read-only service. I'm extremely curious what php.net development would have been like via github (as we were so very close to moving to it instead of SVN) but oh well.
Regardless, it's nice to see a few old timers increase their involvement recently, so now let's hope for some new blood, too.
I was actually glad when the 5.4 branch was created, just because i was hoping i would see php-fpm built-in in a stable release so i could put it live on our production servers asap.
But now im starting to wonder if that will happen at all any time soon after reading this post.
Great write btw, really enjoyed reading it.