About adding "successors" and "precursors" template keywords

[Kostia Balytskyi]

As per Danek's comment, it seems enough to just use 'precursors' and 'successors' revsets.



On 2/19/16, 10:35 AM, "Pierre-Yves David" <pierre-yves.david at ens-lyon.org> wrote:

>
>
>On 02/18/2016 06:24 PM, Kostia Balytskyi wrote:
>> Hi guys,
>>
>> I would like to add "successors" and "precursors" template keywords to
>> the mercurial core. The motivation is clear: to allow people configure
>> their hg styles to display this if necessary (or even tweak the standard
>> mercurial log output if you think this is worth doing).
>>
>> The "successors" keyword will display a plain list of final successors,
>> e.g. for situation like a<-b<-c `hg log –r a` will not display b, only
>> c. What's more, it will display all the successors as a plain list,
>> without nesting. What I mean by that is if commit a is succeeded by b
>> and independently (c,d) (successorssets would return [(b,), (c,d)]), we
>> will just display "b,c,d", without any structure.
>
>The way hgview and thg have been doing that is by showing the next 
>visible successors and that is all. This avoid massive explosion of 
>output size when the chain is long, provide more accurate data when 
>tracking a specific line of meta-history and align with what we do with 
>"parent:".
>
>I would probably starts using this direct successors approach first and 
>see how it goes.
>
>-- 
>Pierre-Yves David