On 11/28/06, Niall Pemberton <[EMAIL PROTECTED]> wrote:
On 11/28/06, Niall Pemberton <[EMAIL PROTECTED]> wrote:
> On 11/28/06, Craig McClanahan <[EMAIL PROTECTED]> wrote:

> > Follow up comment on something I missed (and will apply to Rahul's digester
> > release candidate as well).  Are you planning on respinning the bits with
> > the correct version number (1.3.1 instead of 1.3.1-RC1) before the actual
> > vote?  I prefer to vote on the real bits for an actual release.
>
> I wasn't planning to as its not the procedure[1] Commons is currently
> using. I agree that this is a weakness, since theres potential for
> mistakes when the final release is cut. This needs commons to agree a
> policy change though.

Actually I changed my mind - its only a subversion tag which can be
deleted and re-tagged. I'll do this and upload the proposed final
artifacts to ~niallp @ apache.

Niall

> [1] http://jakarta.apache.org/commons/releases/prepare.html

This is a reason why it's bad:

http://svn.apache.org/repos/asf/jakarta/commons/proper/discovery/tags/DISCOVERY_0_3/

If you tag the RC's as real releases, then there's a large time period
in which a) people can be confused and b) you could lose time and the
release never happen. In the above case that didn't happen - the
release got passed the voting and just wasn't made, but it highlights
why the above would be more dangerous than the current system.

So, -1 to a tag of VALIDATOR_1_3_1 at this stage, and -1 to putting it
in ~niallp/validator-1.3.1/, it should be in
~niallp/validator-1.3.1-rc2/ (or whatever we're on). However +1 to the
file structure you have - we shouldn't be pissing around with 1.3-rc1
in the project.xml etc, we should be just trying for the real release
and then copying those exact files off to the various places.

Hen

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to