Closed RCasatta closed 8 months ago
On my end I get this failure if I try to run the integration tests without modification, but if I comment out two of the tests and run them individually, I get errors related to a hash mismatch.
So I think the minreq stuff is a red herring -- what's actually happening is that we're getting some sort of hash mismatch (which itself is a bug) and then our crazy integration test shutdown logic is confusing minreq.
Basically we should rewrite the integration test harness.
@RCasatta when you run ./run.sh
do the tests pass?
I think the issue is that you're connecting to a mainnet node, so the blockhashes are real blockhashes instead of integration-test ones, so our code asserts on this, then shuts down the HTTP client, then a parallel test panics because the HTTP client is shut down.
Personally I think we should just rip out the paralellization logic and the lazy statics etc etc and just run the tests in series.
@RCasatta when you run ./run.sh do the tests pass?
Yes, and also your analyses is correct (I didn't notice run.sh so I used a local instance).
I am hitting the same error downstream and I thought to have that replicated but here seem another reason
FYI, the error I had in migrating from simplehttp to minreq was due to the fact that minreq requires the protocol to be explicit in the url while simplehttp defaults to http.
From the
integration_test
subdir after setting RPC_COOKIE and RPC_URL