[PATCH] rebase: persist rebaseskipobsolete state

Pierre-Yves David pierre-yves.david at ens-lyon.org
Fri Mar 25 21:06:55 EDT 2016



On 03/24/2016 04:50 PM, Laurent Charignon wrote:
> Yes, but if we don't change the format, we could have rebase --continue
> doing different things depending on what happens between the abort and
> continue.
> I think that we want rebase --continue to show the same things than rebase
> would have showed if it hadn't aborted.

That's the second time you are mentioning that "not recomputing the 
content" can lead to different behavior. Could you provide us with an 
example of such misbehavior to make this clearer to all? (That will 
probably end up in the description of that changeset eventually)

> That goes against the way we do it for rebase normally and I don't feel
> like it is the right choice here.
> Does anyone else have an opinion on that?

Changing the format seems a reasonable way forward. But I would like to 
back you changes with more details about why that format changeset is 
better. The example or misbehavior mentioned above would do it.

Cheers,

-- 
Pierre-Yves David


More information about the Mercurial-devel mailing list