httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject Re: authn/authz split
Date Fri, 30 Aug 2002 15:43:20 GMT
At 10:09 AM 8/30/2002, rbb@apache.org wrote:
>On Fri, 30 Aug 2002, William A. Rowe, Jr. wrote:
>
> > ...  Now that it's GA, we should really be treating the 2.0 tree
> > with the same respect and caution we use on the 1.3 tree.
> >
> > It's time for a 2.1-dev tree, if we want to be playing with new ideas, 
> guys.
> > If they test out clean and don't break compatibility [in any 
> significant way]
> > then they can be backported to 2.0.
>
>The 1.3 tree didn't become "stable" for many releases after the initial
>release.  We definately didn't treat it with the respect and caution that
>we now use until well after the first alphas for 2.0 came out, which was
>years after the first 1.3 release.

Um, I'd like to disagree with you there.  I was told often in early 2000 [prior
to any 2.0 alpha] that "we should save fixing that for 2.0".  Or I would be
questioned, "is that worth a major mmn bump?!?"  The answer was often
no, not for 1.3, and we will deal with it in 2.0.

Don't get me wrong... 2.0 has some bugs to fix.  But we have a dozen coders
with great [new] ideas that need a 'playground' to pursue those ideas.  If they
prove stable and don't break the api, why not backport to 2.0 after the ideas
are proven.

We have enough users who 'played' in the 2.0 alpha tree.  I'm certain some
will be excited enough to stay on the bleeding edge of 2.1.

In the meantime, we've delievered a terrific server, 2.0, that still has a few
wrinkles to iron out.  If we keep breaking the 2.0 tree for "neat new features"
while ironing out those bugs, users will continue to ask when 2.0 will become
stable.  And the answer will be "not in your lifetime."

We took -years- to go to 2.0, because 2.0 is such a massive restructuring.
There is no reason that new features in 2.1 won't be in user's hands by
ApacheCon Las Vegas [at least in beta form.]  Then a 2.2, as long as folks
want to incrementally improve the server, after we call 2.1 golden.

It wasn't until 1.3.14 that we quit flipping around the API.

19980527 MMN bump for the 1.3.0 initial release
19980713 MMN bump for 1.3.1 release
19980917 MMN bump for 1.3.2 release
19990108 MMN bump for 1.3.4 release
19990320 MMN bump for 1.3.5 release

So your statement doesn't hold up that much water, we've gone from the
2.0.35 GA to 2.0.40, bumping several times.  But we should be about ready
to quit it.  Many, many 1.3.6 sources build against 1.3.27-dev.  Many are
even binary compatible.  Somewhere along the way, we quit mucking with
the API.  Maybe we aught to have quit sooner :-)

All this leads to the inevitable question, "when 3.0?"  and that will only be
answered when a few folks put fingers to keys and start to prototype some
radically different async code.  The faux-async models can all be supported
in the 2.x line.  It isn't until we start thread-jumping within a single 
request
that this server architecture just can't cut it anymore.

So... it's time to stop cutting off our developers and start the 2.1 branch or
separate tree.  Let everyone scratch their itches!  I'm happy to see a branch
(leaving 2.0 as head, so nobody following older checkout instructions to grab
the now-current version have a 'surprize' in store.)  I'd also be happy with
a new httpd-2.1 tree.  Doesn't matter one way or the other.

Bill



Mime
View raw message