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.
Bo Berglund wrote: > Now you have got an assignment to solve a problem and you create a > branch off HEAD (possibly after updating your sandbox to get all > the latest stuff). > The solution should be based on current HEAD and while you are > working on this others continue developing on TRUNK. Unfortunately, the TRUNK symbol does not exist and IIRC Tony held the position that we don't need it as it is the same as HEAD (which it isn't!!!). > You want to create a new branch on the BASE of the existing branch. Be careful with your terminology. In contrast to TRUNK, BASE *is* an existing reserved symbol and it does not have anything to do with branches. Rather it refers to the "base revision" of a (modified) sandbox file, i.e. what gets implicitly used as a second argument when you diff with only one -r option. > There needs to be a new flag for update in my view. Something > like: > > cvs update -B -r BranchTag I'd prefer something like this: cvs update -r .BranchTag (or some other symbol pre- or suffixed to the branchname, see my last post) This would IMO be in line with the kind of syntax that can be used with the log command, e.g.: cvs log -r BranchTag. cvs log -r BranchTag,Tag1::Tag2 I believe there are even more valid syntax constructs that could be supplied to comands via an -r option (e.g. for referring to commit IDs) but I couldn't find any comprehensive documentation about this. It's all buried somewhere in the list archives... > Problem: > How to handle files that have been added along the way??? Simple: they did not exist at the time the branch was created, thus they shouldn't be checked out when you request the branchpoint snapshot. Cheers, -- Oliver ---- ------------------ ICQ: 18777742 (http://wwp.icq.com/18777742) MSN: ogiesen at hotmail.com Y!: ogiesen