Closed agallant closed 8 years ago
Now also disabling the pause/resume TCP test for node, which actually nominally passed but cause strange issues where it still ran code after its done()
call (randomly and spuriously breaking following tests). This one is definitely worth more properly revisiting, but for now it'd be helpful to have Shippable/Travis green for node again.
pause resume seems like something we expect to be called and that node should be able to handle gracefully. the cross-instance storage one I don't think we care about either way, so path of least resistance is good there.
Agreed pause/resume should work in node, and it seems to - the bug is that later tests fail, which may have more to do with the test than the implementation (or at least some node-specific quirk with how it runs the test). I'll see if I can debug it but if I can't fix it I'd still be tempted to disable it in node just so the Shippable/Travis alerts become more meaningful.
Came up with a mildly better approach - just ensure the socket is only closed once.
@ryscheng @willscott
Not high priority, but I'm re-disabling this test - I had enabled it because it was passing in freedom(-for-firefox, -for-chrome), but I'm now revisiting freedom-for-node for the first time in awhile (https://github.com/freedomjs/freedom-for-node/tree/soycode-cleanup) and have found that this test does not pass in it.
That, combined with the good question raised by the TODO of whether this behavior is even desired, suggests I should disable the test. However if either of you think this test should be desired behavior (since it is what freedom does in most environments) then I can leave it enabled and focus on getting the node flavor to pass it.