[darcs-users] darcs conflicts/dependencies -- is patch theory the place to start?
AntC
anthony_clayden at clear.net.nz
Tue Sep 18 22:37:41 UTC 2012
Miles Gould <miles <at> assyrian.org.uk> writes:
>
> On 18/09/12 08:06, Stephen J. Turnbull wrote:
> > > Although this sounds alluring from the implementation perspective,
> > > I worry about the user perspective on something like this,
> >
> > bzr does this already.
>
> As does Monotone - see ...
> ... Git has always resisted giving files True Names; Linus'
> reasoning is that renaming files is just a special case of moving code
> between files (see `git blame`'s -C and -M options).
>
So this just changes the question:
- from: "is this the same file as that?"
- to: "is this the same code as that?"
How good is git at detecting same code in different position? How often does
it fail to detect moves (false negative)? How often does it detect a move that
wasn't (false positive)?
The VCS has the usual problem of only intermittently getting a view of the
repo.
I'm thinking we need a hunk move operation, as well as a hunk change.
AntC
More information about the darcs-users
mailing list