[PATCH] rebase: do not raise an UnboundLocalError when called with wrong rev (issue4106)

[Pierre-Yves David]

On 02/13/2014 10:32 PM, Simon Heimberg wrote:
> Am 14.02.2014 01:35, schrieb Pierre-Yves David:
>>
>>
>> On 02/13/2014 04:22 PM, Simon Heimberg wrote:
>>> Am 14.02.2014 01:16, schrieb Pierre-Yves David:
>>>>
>>>>
>>>> On 02/13/2014 03:41 PM, Simon Heimberg wrote:
>>>>> # HG changeset patch
>>>>> # User Simon Heimberg <simohe at besonet.ch>
>>>>> # Date 1392334460 -3600
>>>>> # Node ID 0ed5bb67f3403b8e0655ee5ff2a7aeb32e307837
>>>>> # Parent  0f1ef9e9e904c18f1ac96aef3a0e0d3aa5f1190c
>>>>> rebase: do not raise an UnboundLocalError when called with wrong rev
>>>>> (issue4106)
>>>>
>>>> Is this something happening only when other extension call rebase
>>>> directly?
>>>>
>>>
>>> Exactly.
>>
>> So reading you description, we keep crashing. Was will be the error
>> message in that case?
> No, in this case it aborts telling it can not rebase a changeset on itself.
>
> I should maybe rewrite the commit message.

Yes, my first complain is that I can't decide if the change it good from 
this patch. Adding more details in the description should fix this.

If you patch touch output related things, you want the new output (and 
possible the old one) to be int he description.

>> I tempted to keep crashing early, an AssertionError or
>> ProgrammingError would be fine.
>>
> Maybe this would be better.
>
> Shall I ask to unqueue? And what to choose then?

I already talked to Matt on IRC, I believe it has been unqueued. We'll 
decide what to do next once we have your new commit message.

-- 
Pierre-Yves