<andi->
ahh I see, should follow things more closely again
<samueldr>
there's a workaround where you can build the tests using the (newly introduced?) attribute for it
<gchristensen>
andi-: better than you having to follow it would be it working :P
<LnL>
yeah, I fixed the wrong thing
<gchristensen>
eh?
<LnL>
and I'm not a fan of the proposed workarounds
<andi->
gchristensen: well, I think both are good ;-)
<gchristensen>
me either... I wish it worked like it used to
<gchristensen>
but it isn't helpful for me to just be grumpy about it
<samueldr>
oh, building with `build` here I say is just that, a workaround, and something better should be done, but I'm not in position to look into it :)
<LnL>
I forgot what the specific problem is, but a change basically means you can't run the tests anymore in restricted mode unless you use a magical incantation
<gchristensen>
seems weird, since isn't ofborg's call method the same as hydras?
<MichaelRaskin>
Hydra passes the real git revision ID, I think, but should be close enough…
<LnL>
hydra evaluations result in a similar workaround because of how inputs work
<ekleog>
gchristensen: it's #50342, ofborg isn't passing the `nixpkgs` argument