On Tue, Aug 15, 2017 at 10:15:38PM +0200, Marius Bakke wrote:
> cgit was not affected by the failing git@2.10 package since it only uses
> the *source* derivation.  It does not care about what's going on in
> git's [arguments]. :-)

Ah, that's right.

> So this patch won't change anything from cgits perspective.  I'm
> slightly in favor of keeping git@2.10 until it's really uneeded (awkward
> if we have to add it back later), but no strong opinion either way.
> 
> I would probably have done this instead of disabling tests if I had
> thought about it.

This change won't make a measurable difference to Hydra's workload, but
I'd still like to avoid building this package unnecessarily.