[Python-Dev] Timing for Py2.4 (original) (raw)

Anthony Baxter anthony at interlink.com.au
Mon Mar 29 22:49:13 EST 2004

Raymond Hettinger wrote:

[Anthony Baxter]

Remember: a 2.4 that's broken is far, far, far worse

than a 2.4 that's 6-8 weeks later. Py2.4 is not broken, I use it everyday for everything! It is much more stable than any previous pre-alpha. I'm sensing more FUD than fact. The alpha release is not the same as final release. So, can we compromise and agree to get out a late May alpha but leave the final release date as a range (allowing for your "baking" delay if it turns out that there is some value in letting the bits sit idle for a few months)?

I'm not saying 2.4 is broken. I am saying that there's a pile of large and potentially disruptive changes that are not even checked in yet. At least one of these (decorators) still hasn't had a Guido pronouncement.

Py2.2 didn't get debugged or fully documented for ages. Py2.3 didn't even have a feature freeze until the second beta. This time we will. We're already fairly conservative with two alphas and two betas.

The goal is, if possible, to have as few alphas and betas as possible. If we can kick out an alpha1 that works on a whole pile of systems, and we don't need an alpha2, this is a good thing. A python release is a non-trivial amount of work. I point (again) to the wide variety of systems in the HP testdrive farm, and the SF compile farm, as well as the various systems I have around here. I try very hard to get releases tested and built on as many of these as possible, before I go and cut the release. Saying "oh, just release an alpha1 anyway, if it is completely screwed, we can just release an alpha2 two days later" underestimates the amount of work needed to cut one of these things.

The world is safer with me doing an alpha than with 2.3.4 which has to be perfect. Also, I have no desire to be RM, but that appears to be the only way to avoid months of unnecessary delay.

"Unnecessary" would seem to be in the eye of the beholder.

-- Anthony Baxter <anthony at interlink.com.au> It's never too late to have a happy childhood.

More information about the Python-Dev mailing list