On 2/21/06, Peter Conrad <[EMAIL PROTECTED]> wrote: > [EMAIL PROTECTED]:~/test> tla mv a.c b.c > [EMAIL PROTECTED]:~/test> tla mv d.c a.c > [EMAIL PROTECTED]:~/test> tla commit -L 'test' -- b.c > make-changeset --files: the file specified as b.c > is not the same as the one in the project tree (/test/test/./a.c) > > Oops. One of your claims above must be wrong. (I haven't tried this with > your patch and '-x'.) > OK, the above behavior is still in the latest version. > > What I'm trying to say is actually two things that I shouldn't have mixed > in my previous mail: > > - The problem of partial commits is only partially solved. It would be > nice to have a complete solution. > - I don't see the benefit of your '-x' option. It introduces yet another > partial solution to a known problem, thereby further complicating the > user interface.
It is useful for the user interface. And it was not a solution to the problem of partial commits, but just to improve the usability of the commit command in the common use context (yes, I use the exclude option myself). As you said there are two things here. For the partial commit issue, I already mentioned in another mail to disallow partial commit after a tla mv operation. Only a complete commit can be done after tla mv. This is in line with the aim of making the least changes to the infrastructure in the current state of tla as well as the lack of a clearly "correct" way of handling such cases. Would you think that solves this problem acceptably? > Peter _______________________________________________ Gnu-arch-users mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnu-arch-users GNU arch home page: http://savannah.gnu.org/projects/gnu-arch/
