[PATCH evolve-ext] evolve: do not set experimental.evolution in uisetup

[Pierre-Yves David]

On 04/06/2016 10:27 AM, Jun Wu wrote:
> On 04/03/2016 11:49 PM, Jun Wu wrote:
>> On 04/03/2016 09:01 PM, Pierre-Yves David wrote:
>>> I don't think this will work. There is third party code (especially…
>>> mercurial core) that rely on the extension to set this when enabled. 
>>> And I
>>> assume this would break it all.
>>
>> Thanks. I didn't realize obsolete.py is reading it.
>>
>> How about doing these instead:
>> - evolve.py: remove setconfig
>> - evolve.py: set obsolete._enabled = True
>
> Gentle ping. If we want to prioritize evolve chg tests fixes, we need 
> to have
> a way to get rid of setconfig. I think it's okay to do the above. 
> Opinions?

Unfortunately is a bit too short for me to understand what you are 
actually proposing here. Can you elaborate?

On the mercurial.obsolete._enabled front. Durham did a significant 
amount of work years ago to be able to get ride of it (this created the 
config option and the finer grained selection). However the final bits 
actually removing the boolean apparently never made it. So don't 
remember all the details (and don't have too much setup to do archeology 
right now) but moving more dependency back on obsolete._enabled seems a 
move in the wrong direction (but some archeology is needed and maybe 
finishing killing obsolete._enabled if appropriate)

Cheers,

-- 
Pierre-Yves