[cvsnt] Re: 2.0.58d/mergepoints

David Hauck davidh at netacquire.com
Fri Sep 30 20:41:52 BST 2005


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.


Hi Tony,

Thanks for your comments. I've come to the conclusion (after running a
series of merges and comparing their results) that mergepoints don't help me
with my current situation. In fact there's some pretty strange results in
the mergepoint-reliant test - but I expect this is due to the various merge
history events that are "tagging along" in this merge.

As you mention below the two branch scenario seems to be a natural fit for
mergepoints, and their use, in this case, makes merging quite
straightforward. I haven't used this previously but will now look for
opportunities for its use in the future.

BTW, the symmetric result of the two-branch mergepoint example you describe
below isn't clear to me (as I believe Gerhard is also questioning). This
isn't something that I came across in the documentation, but perhaps this
side effect is explicitly described somewhere else?

Thanks again,
-David

> David Hauck wrote:
> > Hi Tony,
> >
> > I'm currently using the 2.0.58d cvsnt client and server and
> would like to
> > leverage its support for mergepoints. Are there any caveats
> associated with
> > this feature in 2.0.58d that you might alert me to?
>
> IIRC there was an issue with empty branches in that version but it's
> rare to come across it.
>
> > In particular, this time I'd like to leverage the CVSNT
> mergpoint support in
> > order to eliminate the need to specify a common ancestor for
> the merge. I'll
> > be merging from the HEAD into a first-level branch and could
> like to perform
> > this with something like: cvs -q up -k-v -j HEAD. Ideally I'd
> like to merge
> > the head into branch with a fixed tag on the head, but some
> documentation
> > I've seen says that this wouldn't work correctly for merging
> added/removed
> > files. Is this accurate?
>
> There's an ambiguity with merging for removed files (added files are OK,
> normally, since any conflict can be handled in-place).  eg. if you're
> merging branch A onto branch B, and branch A has changes since the last
> merge, but branch B has a remove, then it's a conflict on the whole
> file, which requires intervention (you need to work out whether the
> changes in that file should be kept, copied elsewhere, etc.).
>
> cvs will warn you if/when this happens.
>
> > Another aspect of the branch is that it (branch #1) has previously been
> > received several merges from another branch (branch #2); this
> other branch
> > (#2) has also been merged into the head, but this will be the
> first time the
> > head is merged into branch #1. I'm hoping the use of mergepoints will
> > prevent problems related to conflicts associated with brach #2
> being merged
> > into both branch #1 and the head, prior to the head being
> merged into branch
> > #1.
> >
> A mergepoint is basically point where you say 'for this version Branch A
> == Branch B', and this works both ways.. eg. if you merge a->b, commit,
> then b->a, there won't be anything to merge.
>
> This works well for 2 branches.  For 3 you might have more issue,
> depending on the merge history...  with a little planning you should be
> able to reduce or even remove conflicts from your merges.
>
> Tony
> _______________________________________________
> cvsnt mailing list
> cvsnt at cvsnt.org cvsnt downloads at march-hare.com @CVSNT on Twitter CVSNT on Facebook
> http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt https://www.march-hare.com/cvspro/en.asp#downcvs




More information about the cvsnt mailing list
Download the latest CVSNT, TortosieCVS, WinCVS etc. for Windows 8 etc.
@CVSNT on Twitter   CVSNT on Facebook