Hi Everyone,
I'm sorry guys, nothing personal, but I find the whole discussion a
non-sense if compared with some sane crude reality.
Before all, everyone recognized the "error" of including PM thinking it
would have take 1 or 2 years. This seems to be accordingly to every dev I
heard, the main reason for delaying the beta.
Apparently we are very near that, it looks like the major work should be to
put the infrastructure in the Haiku Inc. servers. Any action, we take to
modify what was planned years ago in terms of compatibility with BeOS will
have the following ensured side effects :
* It will take more development time we have
Making an official x86_64 release will surely enlarge our list of blocker
bugs. I can think myself a few which are not presentable.
Working toward an agnostic ABI will surely take years to become a viable
alternative, while I guess the proposal wasn't directed to R1 itself.
For switching the compat/host architecture similar considerations.
* It will make the Haiku project even less credible than what we are.
Do you remember the PM discussions? Replying to users saying we want to
keep gcc2 for blah, blah, blah reasons?
I'm pretty sure doing something like switching the ABI will destroy Haiku's
credibility. That's just a big chance, to motivately say "What the hell,
are they trolling us?" - "What's that project aim?"
* Future maintainability is surely mined at this point
I used in the last years pretty much only gcc4h builds. I accepted the
compatibility thing for one and only one reason, that it will make me to
confine the current media_kit in the gcc2 compat layer so that I will be
free to rework/redesign it for R2. In either cases we release a 64 bit
Haiku or use a gcc4h system means that I will basically have to deal with
all that until R3.
So I'm still of the hopinion, we have to release a gcc2h beta and nothing
much. The project did already risk a lot after the PM story, do we want to
destroy the project definitely?
I think most users agree that we are already have an R1. But if we want to
be more sage, we could reduce that claim and say "We have a beta, let's
release it."
My overall impression is that there's and there was an enforced
perfectionism in some vital decisional phases of the project, possibly from
people who then left the ball. There's no perfect beta, and there's not
perfect R1, I think we should stop all that unuseful time consuming
discussion and converge *everyone* for the beta.
--
Saluti,
Dario
