Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
On 2004-01-29, Harrison, Andrew <andy.harrison at anite.com> wrote: > > From: Tony Hoyle [mailto:tmh at nodomain.org] > > The extra revision is to record the mergepoint. It's perfectly > > harmless, and only takes a few bytes in the RCS file. > Ok, but why do you need the mergepoint if the file hasn't changed? In the following situation: On the branch: you make some changes. On the mainline: somebody else makes some changes. On the branch: you merge the mainline into your code. You look at the changes from mainline, and realise they are no longer relevant. You remove them. The file you now have is the same as the one you had before the merge. Unless CVS keeps mergepoints around and takes account of them when you do the commit, this code backout that you have performed will not be "remembered". Next time you merge from mainline-to-branch you will have to do it again. Recording the mergepoint and a null delta stops that from happening. phil -- change name before "@" to "phil" for email