Is there any chance of recovery?

[Stephen Rasku]

On Wed, Feb 8, 2012 at 15:14, John Peberdy <johnpeb at gmail.com> wrote:

>
> I think the only real reason rebase should not be shown by default is
> that it is a concept that is not neccessary for beginners to
> understand so showing it is adding unnecessary cognitive strain. I
> don't believe it is dangererous, at least with phases. Would you
> agree?
>

I don't know if I would say it's dangerous but there can be complications
with it.

1. If there are conflicts with rebase it's very hard to figure out how to
resolve them.  Whenever this happened to us we would just revert and do a
merge instead.
2. Unless the developer retests the rebased code then there is code in the
repository that hasn't been tested.  If you merge without testing the merge
a bisect will determine that there's a problem with merge.  However, if you
do a bisect with rebased code than an innocuous checkin could end up being
the culprit which will be confusing to the original developer since he
tested that changeset before it was rebased.

There could be other problems as well.  Rebasing gives the illusion of
working most of the time but when it fails it can fail in a big way.

...Stephen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20120208/48d53c12/attachment-0002.html>