Thanks Paul, but I'm not sure I've got this straight, so bear with me one more
time! :-)
Paul Sander wrote:
>
> You can think of a 3-way merge in the following way:
>
> Given a common ancestor revision A, and two contributor revisions R1 and
> R2, apply the difference between A and R1 to R2. In your case, A is 1.3,
> R1 is 1.4, and R2 is 1.5.
So, R2 will always be the latest revision either from the repository or the
sandbox - is that right? I couldn't understand how CVS knew about R2 but I
think that makes sense now.
Regards,
> --- Forwarded mail from [EMAIL PROTECTED]
>
> Two of our developers were working on the same area of a particular file at the
> same time. The first committed his changes and when the second developer ran
> cvs update, conflicts occurred as you would expect. The second developer went
> about resolving the conflicts but somehow ended up committing only his changes.
> So, we have developer A's changes in revision 1.4 and developer B's changes in
> 1.5 but what we want is a combination of the two.
>
> So, I thought, no problems - "cvs up -j1.4 -j1.5 filename" should merge the two
> revisions. But no - all we end up is the latest revision (1.5). However, if we
> use "cvs up -j1.3 -j1.4 filename" we seem to get the correct conflict output.
> Why?
>
> --- End of forwarded message from [EMAIL PROTECTED]
--
"Everyone is ignorant, only on different subjects." (Will Rogers)
Brian Collins
Triple G Asia Pacific
http://www.tripleg.com
