Mike Hearn
Mike Hearn
BTW, another problem we hit yesterday is running out of DB connections/processes. Same cause; too many tests run in parallel and whilst it should work, the one-pool-per-test-class design means the...
Thanks for the info! My question was whether Windows does any sandboxing at all when `EntryPoint="Windows.FullTrustApplication"`. My understanding was no, and that understanding is re-inforced by what you just said:...
The manifest does _not_ declare `RuntimeBehavior=appSilo` or any other equivalent that would trigger sandboxing. Therefore this perhaps isn't a question at all but a bug report. Something appears to change...
@DrusTheAxe I haven't reported it because I don't have access to the app that had the problem, I am a middleman unfortunately. If I find a way to reproduce I'll...
It was just an example, there's another failure after that. What version of Java do you compile it with?
@graemerocher @dstepanov I've added a test, avoided changing the existing OrderUtil API and skip the sort if the list isn't big enough to need one. I also renamed the test...
CI failure looks like a flake rather than a result of this change. I see quite a few flakes locally related to the HTTP/Netty tests.
Everything seems to pass except the flake I already fixed in 4.9.x, so once it's merged up that test would have passed too. I think this is ready for merge.
@graemerocher It does target 4.10.x now (I rebased and retargeted the PR already).
I tweaked the test. Note that the reason this is so complex is that event listener dispatch works in an odd way - it has a `List