Re: Patches

[Randy Terbush <ra...@zyzzyva.com>]

> Jim Jagielski writes:
>  > As a general trouble maker, let me see if I have this right.
>  > 
>  >    Patch A: Changes things like HAS_GMTOFF to HAVE_GMTOFF and is
>  >     derided loud and long as being too much too late. Patch
>  >     removed.
>  > 
>  >    Patch B: Involves some heavy-duty changes to mod_cern_meta.c
>  >     as well as a semi-substantial change to the way the
>  >     module works (from per-server to per-directory). It is
>  >     added "much" later that Patch A. The resultant outcry?
>  >     One very nice message stating "isn't this kinda big?"
>  >     Patch remains.
>  > 
>  > 
>  > Is something wrong here?
> 
> Fair point in my opinion. Are we actually in code freeze now? This last
> patch went in days before a release is supposed to be cut, exactly how
> much testing is it going to get?


At risk of getting the greased pole.....

The point that both I and Jim raised when this issue last appeared was
the undoubtedly "longer shelf life" that 1.1 will likely see. No one
commented at all about this point which I think is valid.

Why can't we branch the CVS tree and get started with 1.2?
Wasn't this the advantage of setting up CVS?

I think that we are still seeing valid changes and cleanup patches
coming in like Andy and Jim's and would like to see them go into
the 1.1 release. By branching and waiting for things to die down
a bit on 1.1, we stand a much better chance of not repeating the
1.0 release scenario. The reason the 1.0 scenario happened was that
we got in a big toot to shove it out the door.

Re: Patches

[Paul Richards <p....@elsevier.co.uk>]

Randy Terbush writes:
 > At risk of getting the greased pole.....
 > 

greased poles are fun :-)

 > The point that both I and Jim raised when this issue last appeared was
 > the undoubtedly "longer shelf life" that 1.1 will likely see. No one
 > commented at all about this point which I think is valid.

It is valid and would be acceptable if the release process was better
though out but we're too close to getting it out the door to discuss 
release procedures for 1.1, let's get it as stable as we can and postpone
the discussion until next week when 1.1 is over and done with.

 > 
 > Why can't we branch the CVS tree and get started with 1.2?
 > Wasn't this the advantage of setting up CVS?
 > 

I advocated this in the first place.

 > I think that we are still seeing valid changes and cleanup patches
 > coming in like Andy and Jim's and would like to see them go into
 > the 1.1 release. By branching and waiting for things to die down
 > a bit on 1.1, we stand a much better chance of not repeating the
 > 1.0 release scenario. The reason the 1.0 scenario happened was that
 > we got in a big toot to shove it out the door.

This isn't always the solution since a lot of the developers go off
into the -current branch and play with new toys and no-one does the
work to get the release polished up. Since we're going to do a release
imminently I suggest we just carry on for the next week. I plan to
branch and tag the tree when 1.1 is done so we can work this way next
time. I'll post a proposal on how I think this should work once 1.1 is
out the way.