Alex Rousskov wrote:
On Thu, 2008-02-21 at 14:46 +1300, Amos Jeffries wrote:
- include directive. Being a very minor feature, and existing in 2.x. I'm
undecide if it should be back-ported early to ease future upgrades from
2.6/2.7.
I would be happy to see it backported to 3.0. It does not affect runtime
code and does not affect old configs, so it should be safe.
Eventually, things like that would go into the next minor release (i.e.,
v3.1 in this case). Currently we have too many disruptive features going
into minor releases and cannot release them often enough for the "keep
the subreleases free of new features" scheme to work well.
what do you think of habitually considering back-porting where
possible, the features added to 3.1 that are actually forward-ports from
2.x?
Sounds like a good idea to me, _considering_ being the operative word.
We do not have a list of 1000 features to go through, so it does not
have to be rigid and formal. We can decide on a case-by-case basis.
You catch my meaning exactly. Great.
I'm thinking some basic criteria in a 2-minute thought process along the
lines of:
Did it exist in 2.6 before 3.0? 2.7 before 3.1?
- No => probably ignore.
Is it just a patch-and-tweak (not a lot of work)?
- No => oh well
Is it a big change to existing code (_new_ code gets past this)?
- Yes => oh well goodbye.
Is it worth a few minutes my time?
- No => oh well, goodbye.
** Okay, it _might_ be worth it. Spot-check core for vetos.
Amos
--
Please use Squid 2.6STABLE17+ or 3.0STABLE1+
There are serious security advisories out on all earlier releases.
