Stephen Leake <[EMAIL PROTECTED]> writes: > Richard Riley <[EMAIL PROTECTED]> writes: > >> Not having the branch commands in DVC for git would immediately >> invalidate its usage for a lot of people I think. > > That may be true, but I hope not.
My own gut feeling is that it is. And current DVC uptake by git users also indicates that - every time I have asked people say "use magit" which is stage and branch orientated. Especially when "competing" packages do perform those features. See magit for example. The research I have done recently looking at the competing packages shows that one of the reasons people dont use the otherwise good builtin emacs 23 VC is that it does not support branching. I only found out about DVC from the emacs irc channel and Stefan has since updated this page to include a pointer to DVC: http://www.emacswiki.org/cgi-bin/wiki/Git so that more people will be aware of its presence. > > Every GUI interface to a CM system I've used requires occasional > command line use. For some operations (like renaming lots of files), > the command line is better. > > I use the commannd line for branching in mtn; there is no xmtn-branch > command in DVC. Personally I think needing to go to the command line indicates a functional short coming. Yes, one can. But I am only suggesting improvements to help DVC become more widespread. It is closer than the rest to a generic front end I think. > > In fact, I wrote a customized mtn command for branching, because I > have four related projects that should be branched together for my > simulators (named GDS); makerules (a collection of core make files), > sal (a low-level library for spacecraft math), common (the core shared > code for the simulator), lro (the Lunar Reconnaisance Orbiter > mission-specific simulator code). there are also other missions. Specialised projects will surely need specialised commands but standard branching is a relatively "must have" requirement for any git interface I would have thought. > > mtn commands can be written in Lua, and invoke other mtn commands. In > this case, it was easier to do that, and the resultant command runs > faster, than doing the equivalent in elisp for a custom > xmtn-gds-branch command. I suspect most things run faster than in elisp. But 0.02 seconds or 0.001 seconds don't make much difference for most users on an action invoked from a hot key. If run time was an issue a lot of elisp features would come last in the efficiency stakes :-; > > That said, there may be a reasonable way to implement a dvc-branch > command. What would you like it to do? The same as the command line. In fact Stefan has implemented it already. The select branch command seems to have a small bug though which I reported to him last night. Your own tutorial talks about branching. Or was this only for the status report? http://gds.gsfc.nasa.gov/dvc-intro.html#Status-Display > >> branching and staging are two of the most used features of git in my >> (limited) experience. > > Once you get going, commit and merge will be more common; you should > not have to create branches often. Branches are very common I think. Most people branch for individual fixes and when tested merge back in. In fact nearly all the tutorials I see for GIT centre around the branch, work, commit, merge paradigm. See the individual developer workflow here: http://www.kernel.org/pub/software/scm/git/docs/everyday.html regards Richard. _______________________________________________ Dvc-dev mailing list [email protected] https://mail.gna.org/listinfo/dvc-dev
