[PATCH 4 of 5 relative-diff] patch.trydiff: add support for stripping a relative root

[Siddharth Agarwal]

On 03/17/2015 04:43 PM, Martin von Zweigbergk wrote:
>
> Too costly to check that they really do? I saw some patches from Augie
> about [devel] config. Would that make sense? Or it's just too unlikely
> to break so it's not worth it? I would accept that answer. It's just a
> little funny that the 'relroot' content is ignored. I'm sure you
> reflected on this too.

Reasonably unlikely it'll break -- this function is very low level and I
can't think of anyone that would want to use it. Adding devel checks
(and turning them on in tests) makes sense I think.