Bazaar still below the radar when evaluating VCS tools

Stephen J. Turnbull stephen at xemacs.org
Fri Feb 26 07:50:44 GMT 2010


Ben Finney writes:
 > "Stephen J. Turnbull" <stephen at xemacs.org> writes:
 > 
 > > Andrew Bennetts writes:
 > >
 > >  > I'm going to guess though: I think Stephen might be saying that none of
 > >  > the systems automatically capture a 'mv foo bar' done in their working
 > >  > trees, when ideally they would.
 > >
 > > Precisely.
 > 
 > In that case, I disagree :-) It's not necessarily the case that the VCS
 > should automatically pick up on that, and I prefer that it not guess.

As long as you actually do use system functions rather than bzr, you
don't really have the choice of preventing the VCS from guessing.  You
can have it guess the constant "it's not a copy/move," and like the
proverbial stopped clock it will be right some of the time but you
won't know when.  Or you can have it use some more accurate heuristic,
in which case it will be right more often but you'll be more surprised
when it's wrong.

We could of course prevent bzr from guessing; if there are *any*
changes (including ordinary edits of a single file) that were not done
via bzr cp or bzr mv, you could have bzr interactively query whether
any of them are copies or moves.  This would get old *real* fast.

The real alternative that you have in mind is to discipline yourself
to use bzr when you intend a change that should be recorded in the
VCS, and to use system commands when you don't.  That's a reasonable,
even admirable, position, but you shouldn't be surprised that others
want the VCS to adapt to their less precise workflows.




More information about the bazaar mailing list