[PATCH 1 of 2] discovery: weaken claim about returned common heads if ancestorsof are given

[Manuel Jacob]

On 2020-07-17 08:49, Pierre-Yves David wrote:
> On 7/17/20 6:16 AM, Manuel Jacob wrote:
>> On 2020-07-17 05:15, Pierre-Yves David wrote:
>>> This looks correct to me. Thanks of lot to Manuel for making this
>>> clarification work.
>>> 
>>> You should consider using mercurial.util.nouideprecwarn to catch (and
>>> have extension catch) user of the old attribute.
>> 
>> In the current patch, I forward the new name to the old name without 
>> changing the users. However, I think that changing the name at all the 
>> users is a good idea because then it will become more obvious in the 
>> code if there is a bug. We’ll need to review all places using the 
>> attribute (and possibly fix some). Using "ancestorsof" at all the 
>> users will make the fix more obvious in the patch.
>> 
>> I could either:
>> 
>> 1) send a follow-up patch doing the mass-rename and adding the warning 
>> (this means the already sent patch stays less noisy), or
>> 2) send a new patch combining the docstring change and the 
>> mass-rename.
> 
> Sending follow ups seems better.

I send such a patch in reply to the patch fixing the docstring.