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.
kerry richard wrote: > Howeever, if User2 attempts a 'cvs edit' on the Branch1 version of the > file, it succeeds ! This I find surprising..... I think this is wrong as > it means anyone finding a file is locked for editing can get around the > lock just by requesting to edit the file on a branch. As the reason we > want to use locking is for files that can't be merged, this seems to be > a problem. Ie the file can't be merged so we're enforcing locking, but > the lock can easily be circumvented using a branch which gets us back > into the position of a merge being required. > > Am I missing something ? If you really don't want to merge the files, you shouldn't branch them. This is the whole issue. Once you branch them, you may have to merge them (or decide which one to keep). Say you use the global 'cvs edit -w', as Tony says. What good does it do for you if User1 'cvs edit -w' the file on HEAD, then commits it, then User2 'cvs edit -w' the file on Branch1, then commits it? From the 'cvs edit -w' point of view, everything is ok. But in the end, you still will have to either merge them or decide which version you want to keep -- this is completely independent of whether the edit was global or branch-only. Gerhard