[PATCH 6 of 6] [RFC] extension: `dummy` and composition with `lfs`

[Augie Fackler]

> On Nov 16, 2016, at 09:12, Rémi Chaintron <remi at fb.com> wrote:
> 
> On 11/9/16, 5:17 PM, "Augie Fackler" <raf at durin42.com> wrote:
> 
>>    On Thu, Oct 27, 2016 at 04:04:04PM +0100, Remi Chaintron wrote:
>>> # HG changeset patch
>>> # User Remi Chaintron <remi at fb.com>
>>> # Date 1477579974 -3600
>>> #      Thu Oct 27 15:52:54 2016 +0100
>>> # Branch stable
>>> # Node ID c7dc98eb24cd40e46e73dea275b4c5b245155efb
>>> # Parent  202b2a6872d3b3ffb1f373910c4eadffda8f727a
>>> [RFC] extension: `dummy` and composition with `lfs`
>>> 
>>> This is a proof of concept for the composition of non-reflexive operations on
>>> filelogs using extensions relying on the flagprocessor design.
>> 
>>    Could this be tested using censor instead?
> 
> I currently have a prototype to move various bits and pieces from censor related core code into the extension, but given the extensive rework and redesign involved, I found it more applicable to provide an easier proof of concept instead.
> I’d love to get a proposal up for review in regards to moving censor to the extension, but I’m not positive I will have the time to get everything in order to make it happen before the end of the year.

Does that mean that with this series as-is, censor doesn't use the new code paths? Have you explored how much refactoring has to happen in censor to make censor use the new system?

I'd rather that the first-party consumer of this infrastructure used it right away, rather than it being a TODO unless it's extremely onerous.