Proposal on how we should handle syncing between stabilisation forests and always open 7u

[Chris Hegarty]

On 04/11/2011 09:58, Alan Bateman wrote:
> On 03/11/2011 16:02, Edvard Wendelin wrote:
>.....
>>
>> If you have any feedback, please let me know before Monday.
> Edvard - what does this mean for the history? Just thinking of someone
> with a clone of jdk7u/jdk7u, they will no longer be able to do a "hg
> update" to update the working copy to get to any build of any update,
> they will only be able to get to builds that were done prior to forking
> for stabalization. Also just wondering if it will be confusing to have
> different changeset ids for the same fix.

Good questions Alan (which I don't have the answers to!)

I just wanted to point out that the jdk7u-dev forest of repos allows 
duplicate bugIds's, and some developers have already been pushing the 
same changes into both the stabilization forest and the jdk7u-dev forest 
using different changeset ids. I don't like this approach because of 
what it does to the history/merges/etc.

I'd be more in favor of an auto sync (or something) between the 
stabilization forest and jdk7u-dev. Or where possible import the exact 
changeset id into jdk7u-dev directly after the push into the 
stabilization repo.

-Chris.

>
> -Alan.