/hg/release/icedtea6-1.5: Added tag icedtea6-1.5.3 for changeset...

[Mark Wielaard]

On Wed, 2009-09-02 at 10:36 +0100, Andrew Haley wrote:
> Mark Wielaard wrote:
> > On Tue, 2009-09-01 at 23:15 +0100, Andrew John Hughes wrote:
> >> 2009/9/1  <langel at icedtea.classpath.org>:
> >>>        Added tag icedtea6-1.5.3 for changeset d0593233c9ce
> >> Huh? Has there been a 1.5.2 release? I didn't see an announcement.
> > 
> > Although it is technically possible to move a tag after it has been
> > created, I don't think it is a good idea. Allocating a new tag name
> > seems a better choice.
> 
> Not to me.  We surely don't want phantom tags for nonexistent releases.
> This is very bad practice; IMO, as usual.

OK, but then we need to better communicate when a tag/release is made. I
had assumed that a tag is only created when the intention is to declare
that changeset as the release. Tags that aren't stable feel wrong (IMO
also of course), so 1.5.2 was tagged, but deemed not a great release,
lets just better ourselves with the 1.5.3 tag/release. But I now assume
that assumption was wrong :)

So may I humbly suggest that even "routine" commits like these get a
message to the list that explain their intention ("yeah, I am tagging
this now as release-x-y-z, but don't count on it yet, it might move
later").

Cheers,

Mark