<abathur>
can you check if NEXT_ROOT is in the environment at that point? I see some similar reports that say it wasn't, but it is a fairly common theme in many of them it seems
<supersandro2000>
and the system log contains so much noise that I can barely find anything
<supersandro2000>
I think it tries to do a network bind which would explain the failure in sandbox
<supersandro2000>
but I allowed that already
<supersandro2000>
whatever. I just disable them for now
<abathur>
how do you know it's that one? just manual testing (didn't see it mentioned in the log paste)? Do you know if that breaks when it imports cocoa, or on a specific line/op in here?
<abathur>
also, it's clunky to come up with the predicates, but once you do the "log stream" command can be a fairly tolerable way to monitor the log firehose
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<supersandro2000>
did not copy it in
<supersandro2000>
but it is the only thing that is not spammed for every test
<supersandro2000>
it breaks on a specific test
<supersandro2000>
also the tests are super unstable in sandbox
<supersandro2000>
why do I even care and not just disable them?
<abathur>
as in, flaky?
<supersandro2000>
fails 40% of the time
<abathur>
fun
<abathur>
flaky won't do, so I guess they either need to get patched locally, fixed upstream, or disabled
<supersandro2000>
I won't patch the tests except if it is very easy.
<supersandro2000>
Upstream fixes are unlikely because the dev did not react to my one month old issue with a question
<abathur>
I can think of one case, where maybe the code (or the tests) have optional dependencies on something external, and they're finding it without sandbox, but not with, and falling into alternate codepaths
<abathur>
maybe diff the build outputs with/without sandbox in case you can see anything it's optioning or falling back on?
xcmw has joined #nix-darwin
<abathur>
yeah, I'm not sure; I know I've noted before that the contributor base is very narrow, but that it has a lot projects directly or indirectly dependent
<abathur>
only 3 people with more than 8 commits and only 5 with more than 2
<abathur>
I'd base that call more on how he handles PRs than issues