V1.0

[Ben Laurie <be...@gonzo.ben.algroup.co.uk>]

I have been thinking about the 1.0 question (who hasn't?). It seems to me that
we have two conflicting desires, firstly, to have a stable release, and
secondly, to just get 1.0 over with so we can get on with the interesting bits.

The current release cannot considered to be stable, with bug reports averaging
at least 1 per day, but it is obviously getting there. I think moves towards
labelling any release as 1.0 without it being out in the field for at least
a few weeks with no serious bugs reported are doomed to failure (mostly through
passive resistance). However, in the meantime, none of the new work can be
done. There is a simple solution to this: branch 0.8.16 to 0.9.0 (or 1.1.0),
maintain 0.8.16 as a release candidate, and let us get _on_ with it on 0.9.0,
fixing bugs in 0.9.x in parallel with 0.8.x.

I proposed this before, and met with silence (like many other 1.0 proposals).
If people don't like this idea, could they at least say why, this time?

I am keen to get the structural changes required to make SSL a light patch
instead of a heavy one (though making it a module would probably fall afoul
of the crypto hooks biz), and also to do the code cleanup which has been
threatened for so long (you should see how many warnings I get at warning
level 3!).

Hey - I just had a neat idea - if I frame this proposal the other way round
(i.e. I propose that we don't make a branch version) and then veto it, by the
rules this means that we must make a branch version, doesn't it?  :-)

Cheers,

Ben.

-- 
Ben Laurie                  Phone: +44 (181) 994 6435
Freelance Consultant        Fax:   +44 (181) 994 6472
and Technical Director      Email: ben@algroup.co.uk
A.L. Digital Ltd,
London, England.