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.
A more simple example, we have tested: - FileA has a branch "release-1_0-branch" - development goes on and due refactoring FileA is renamed to FileB (or moved to somewhere else in the project) - now a bug is reported for version 1.0 and it needs to be fixed in release-1_0-branch - trying to merge the changes back to the main trunk looses the made changes - CVS must recreate FileA in the maintrunk and mark it as conflict, otherwise the changes are lost!!! => then we can merge the changes manually to FileB --Mike On Wed, 3 Sep 2003 10:02:50 -0700, Glen Starrett <grstarrett at cox.net> wrote: >> Well, deleting the file is nearly the same as removing all >> lines from it. > > My $0.02.... Looking at this from a more code-oriented view, let's walk > through more specific examples. > > For all example, assume File A is sourcecode that uses File B. Perhaps > Dev 1 edited File A and removed all use of File B, then deleted File B > to be tidy. This was done on MAIN. > > ** I haven't tested these, but I believe it works this way ** > > > Example 1 > ========== > Now along comes Dev 2 on their own branch. They modify File A to use a > new function in B and change B to add that new function. > > Merge Dev 2's branch into HEAD, and you WANT File B to be there, since > it is dependent. > > How would CVS know to remove B? It was changed, it needs to be merged > in, it is part of the valid sandbox on the branch that you want brought > into the MAIN. > > > Example 2 > ========== > Dev 2 on a branch edits some files, but NOT B. Since B didn't change > from the merge point, it is not merged into the sandbox and is not > re-added to the HEAD. > > > Example 3 (possible solution) > ============================== > Dev 2 is on a branch and makes some changes to any files out there. > They know they must merge their changes onto MAIN. > > They perform the merge on their branch, by merging HEAD to their sandbox > FIRST. The irrelevant File B is removed from the branch at that point. > > Now, when the dev merges to HEAD, File B stays away and everyone is > happy. I think. > > > Is this correct? Would example 3 work the way you want it to? With > cvsnt's mergepoint processing making this easy I've been doing that > scenario here and it seems to work very nicely, although I haven't tried > it yet with file removal. > > Regards, > > Glen Starrett > > > -- Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/