Re: bzr repository ready?

[tomas@tuxteam.de]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Fri, Nov 27, 2009 at 08:22:31PM +0100, Lennart Borgman wrote:

[...]

> Hum. That seems pretty important even though I am pretty sure I do not
> understand all the details.

The principle is quite easy (although details can get messy, as always):
if you have other people depending on your (published) repository, you
better always "move forward" and don't mess with the past (and rebasing
is a mild way of messing with the past). Otherwise "your" past and "your
client's past" won't agree, and that isn't funny.

> But why are not normal human beeings protected from the evil of
> rebase? Why do they have to know about it?

Because when you are developing a local patch derived from some
upstream, rebase is too convenient to ignore it. It lets you "see" your
changes always relative to the "current upstream version" ("as if" you
started developing from there -- and that's the "history changing" bit).

To sum it up -- you may practice revisionism whenever nobody's looking.

Regards
- -- tomás
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFLEMcTBcgs9XrR2kYRAnjjAJ40E+aDcLAO99oc5E9P7m2L+9n+HACfbFrH
De3zaiZBC4HnjZRbR70mZRg=
=OVZ4
-----END PGP SIGNATURE-----