Aaron Bentley | 1 Feb 2005 21:32
Favicon

Re: Partial update

Jan Hudec wrote:

>> star-merge is superset of update.

I used to think so, but it's not accurate when you look at trees that 
have deliberately diverged.  Star-merge will throw away local changes 
that were rejected by the other branch, and merge the two as closely as 
possible.

Update, on the other hand, will retain local changes, and apply any 
changes from the other tree since the last merge from other.

This is why Fai supports "merge --update".  It also supports "merge 
--update --diff3".

Aaron

--

-- 
Aaron Bentley
Director of Technology
Panometrics, Inc.

Jan Hudec | 1 Feb 2005 22:10
Picon

Re: Partial update

On Tue, Feb 01, 2005 at 15:32:18 -0500, Aaron Bentley wrote:
> Jan Hudec wrote:
> 
> >>star-merge is superset of update.
> 
> I used to think so, but it's not accurate when you look at trees that 
> have deliberately diverged.  Star-merge will throw away local changes 
> that were rejected by the other branch, and merge the two as closely as 
> possible.
> 
> Update, on the other hand, will retain local changes, and apply any 
> changes from the other tree since the last merge from other.

Well, I am not sure how update behave in presence of skipped
revisions. Star-merge <tree-version>-somepatch will not pull them in.
Replay will. Without skipped revisions, they should do the same (apply
diff between last applied patch and the requested one).

> This is why Fai supports "merge --update".  It also supports "merge 
> --update --diff3".
> 
> Aaron
> 
> -- 
> Aaron Bentley
> Director of Technology
> Panometrics, Inc.
> 
> 
> _______________________________________________
(Continue reading)


Gmane