Workflows, rebase, patch theory
Ben Finney
bignose+hates-spam at benfinney.id.au
Wed May 7 13:18:36 BST 2008
(Please preserve attribution lines on quoted material when replying,
so that we can see who wrote what in your reply.)
Talden <talden at gmail.com> writes:
> > > AIUI, git users tend to use this to keep their changes up to
> > > date and "ready for submission" with respect to a trunk. The
> > > advantage is you don't have commits like "Merge bzr.dev"
> > > interspersed with your own changes.
> >
> > Okay. This does seem to be a main complaint against Bazaar and
> > Mercurial by Git or Darcs users. I don't understand why it's
> > desirable to hide merges in the history, especially against...
>
> Well I guess because in a longish running branch that has high
> dependencies with other branches landing in the upstream you can get
> a lot of merge-noise.
I don't see how "information about what happened when" is "noise",
especially in a system that's designed to preserve exactly that
information — which is what a VCS is designed to do.
> In an ideal world all tasks would involve a handful of commits over
> a day or two and not require these upstream merges... Not all
> projects fit this model.
Perhaps. So, in non-ideal cases, it seems folly to throw away
historical information about what actually happened, even if it's less
than ideal.
--
\ “As far as the laws of mathematics refer to reality, they are |
`\ not certain, and as far as they are certain, they do not refer |
_o__) to reality.” —Albert Einstein, 1983 |
Ben Finney
More information about the bazaar
mailing list