Re: [gentoo-alt] Restructuring the bootstrap (or: getting rid of gcc 4.2)

[Ruud Koolen]

On Monday 03 February 2014 02:23:12 heroxbd@g.o wrote:
> Hey Ruud,
>
> I'll omit your explanations which I am aware of: It's a nice wrap up
> indeed. So I'm going to cut in here.
>
> The new stageX feels intertwined:
>
> stage2: tmp portage
> stage3: tmp gcc, EPREFIX gcc, EPREFIX portage
>
> How about defining stage2 as "anything to be built with tmp portage into
> tmp", and stage3 as "anything to be inserted into EPREFIX"? Then it's
> arraged as:
>
> stage2: tmp portage, tmp gcc
> stage3: EPREFIX gcc, EPREFIX portage

The reason I haven't done this so far is that many of the temporary portage 
settings (bootstrap USE flags, FEATURES, etc) apply equally to both parts, 
and it isn't obvious how much duplication of code this would entail. But I 
plan to look at this in more detail once the main bulk of the work is 
accepted and stable, as to not make the atomic change set larger than it 
needs to be.

> On the resistence from the momentum of tradition, I hope you can look at
> the situation from Fabian's angle of view: the fragile script (bugs
> sometimes) at least works on AIX, HPUX, solaris, MacOSX. They might be
> broken by the new bootstrap concept. As you don't have access to these
> platforms, several people have to be involved to prove the platforms are
> cool with the new concept, which implies a lot of work and social
> coordination than you might anticipate.

Quite. I'm trying to do as many platforms as possible by myself, but there's 
no avoiding that a change of this magnitude will involve contributions (at 
least testing, and possibly debugging) from several different people.

> Unless the present solution is a real pain in the ass, we are conservative
> with the refactorization. The new one is good enough and it's a good job,
> while it's counterpart should be bad enough to drive your fellows on track.

As I explained, I do think the new method solves real problems, and will make 
supporting RAP reasonably possible without maintaining two independent pieces 
of infrastructure. I agree with being conservative, but I am kind of allergic 
to the idea of holding back useful progress *indefinitely* in the name of not 
fixing what isn't broken; there must be SOME mountain of due care large 
enough to overcome that consideration.

> If, by all means, you got development boxes of AIX, solaris, etc. And
> prove the new one can 100% replace the old one. That's when it starts
> getting widely accepted. At the same time, however, this process will
> take a lot of work. If it is your destiny, go for it.

I started this thread because I expect[ed] to be at the point where getting 
things to work on all architectures should be reasonably straightforward. I 
think the only step remaining is to actually go and test it, dig in, and see 
what problems emerge [no pun intended]. Hence this thread, of course.

As the champion of this idea, I'll gladly do the work to make sure the new 
bootstrap is solid on as many platforms as I can get access to. The more, the 
merrier :)

-- Ruud