Weird failure of a merge to local branch

[Eli Zaretskii]

> Date: Thu, 09 Dec 2010 14:56:35 -0500
> From: Aaron Bentley <aaron at aaronbentley.com>
> CC: bazaar at lists.canonical.com
> 
> Here are some of the steps in a pull:
> 1. Update the repository with new revisions
> 2. Update the branch tip to point at a new revision
> 3. Update the contents of the working tree.
> 4. Update the working tree tip to point at the branch tip revision

May I suggest that this be added to the docs?  This is the kind of
information that is extremely helpful when one gets in trouble and
needs to understand the semantics of the relevant bzr operations, to
figure out how to get out of the woods.

Also this description of corresponding steps in "bzr update" would be
very useful in the docs:

> It updates your tree tip to match your branch tip, and updates your tree
> contents to match.

And while we're at that: what would be the corresponding description
of "bzr revert"?  Is it true that it does step 3 above, touching
neither the branch tip nor the working tree tip?

> When you use [bzr update] with a bound branch or heavyweight
> checkout, it behaves similarly to a pull.

What, if any, are the differences between pull and update in this
case?

Thanks.