> Hi Tanya, > > I do not really understand this strict rule :-) > While I get it from a CVS perspective, in Subversion > this is simply overkill. For making a branch you simply > select a revision X, do a "svn -rX up", test, > if successful, you do "svn info ." and use the URL > to make a repo -> repo copy: > > svn cp -rX URL BRANCHURL > > that's it. Noone can spoil your work.While you are correct is is not usually a problem, that assumes that I do not check in any changes. If I check in X and someone checks in something (Y) that should not go into the release, and then I check Z for the release, I then have to deal with getting Y out. Not a big deal, but I really want to make this simple. I don't really think 15 minutes of people not checking into SVN is that big of a deal. I don't even do anything on the server side, I just ask people to be kind. -Tanya
On May 9, 2008, at 12:16 PM, Tanya M. Lattner wrote:> > While you are correct is is not usually a problem, that assumes that > I do > not check in any changes. If I check in X and someone checks in > something > (Y) that should not go into the release,If they are checking in something, trivially, they want it to go into the release. If they didn't want it in, they'd wait until after you said you created the release branch... ?
On Fri, May 9, 2008 at 5:09 PM, Mike Stump <mrs at apple.com> wrote:> On May 9, 2008, at 12:16 PM, Tanya M. Lattner wrote: >> >> While you are correct is is not usually a problem, that assumes that >> I do >> not check in any changes. If I check in X and someone checks in >> something >> (Y) that should not go into the release, > > If they are checking in something, trivially, they want it to go into > the release. If they didn't want it in, they'd wait until after you > said you created the release branch... ? >And of course no one ever makes mistakes and commits things accidentally. :-) -bw