New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make changes to make all of the test jobs green #1981
Conversation
0881a1f
to
de66063
Compare
de66063
to
57cbf5c
Compare
The remaining Mac failure is bizarre. The failing test is to have the DNS resolver resolve the target name EDIT: Wikipedia has the answer:
|
bec6fea
to
4d89192
Compare
0e58f6a
to
770c50c
Compare
770c50c
to
8d77104
Compare
ec3b662
to
2af9a05
Compare
This skips three groups of tests:
NoCompress
flag test. We currently use theencoding
argument of thewrite
method to pass write flags in. This is not supported functionality, but it has worked for a long time. However, Node 16 apparently added validation for that argument, and any flags we pass in are not considered valid. So, that now does not work at all, and as a result we cannot pass in write flags at all. We will need to find a new way to do that, but for now we are skipping the test of the way it would work currently.loopback4.unittest.grpc.io
andloopback46.unittest.grpc.io
. In the first case, since there is no IPv6 address, the Mac DNS resolver returns an IPv4-embedded IPv6 address, which is out of the scope of the test to test for. In the second case, that address should resolve to127.0.0.1
and::1
, and it works correctly on Linux. However, on Windows we only get the IPv4 address and on Mac we only get the IPv6 address, so there is nothing we can consistently test for on all platforms.That revealed an actual bug: with the introduction of compressed message handling on the server, the message handling code now performs an asynchronous operation, which can result in reordering of messages, or messages with the end of stream event. The second commit corrects that.