--On Thursday, August 12, 2004 2:10 AM +0200 André Malo <[EMAIL PROTECTED]> wrote:
There are also some showstoppers in 2.1 which I don't see resolved until Oct 1 or Nov 1. IIRC Nick had also a kind of roadmap (?) posted some weeks ago.This one should also reviewed.
Well, progress isn't going to happen by just talking about it. People need to come up with code patches. ;-)
I honestly don't think we need to wait for any features or whatnot. Place a hard deadline: if you don't make it by October 1st in *and* it's not an agreed-upon 'release showstopper' in STATUS, then too bad: wait for 2.4. (How long has my mod_auth rewrite been sitting in 2.1? *sigh*)
As a better way I think, we should start to write down, what our plans for 2.2 really *are*. Just to say "feature freeze" is not really enough (and experience shows, that it doesn't work)
My short-term goal for 2.2 is mod_cache out of experimental. It's probably close enough now (at least mod_cache and mod_disk_cache) that I think we could move it out now with the intention of fixing up the remaining bugs and RFC violations in time for any 2.2 cycle. (mod_mem_cache strikes me as being way too complicated and with a lot of potential for subtle race conditions.)
My biggest problem is a lack of access to networking equipment: which I'm in the process of resolving (ever so slowly). The people who already seem to have the network don't want to run HEAD snapshots, so that sort of defeats my intention of getting mod_cache running even faster for 2.2. However, I think it may make sense to put the performance aside for a bit and focus on the RFC compliance in mod_cache for a bit.
I really think the mod_auth stuff makes it sooooo much easier to write aaa backends, that we're doing a great disservice to our module writers by holding back on 2.2 even by a single day. ;-) -- justin
