Draft: client: Finish all opentracing.Spans in test
We've been experiencing flaky tests in client/dial_test.go [1]:
=== RUN TestDial_Tracing
=== RUN TestDial_Tracing/unary
=== RUN TestDial_Tracing/stream
dial_test.go:355:
Error Trace: dial_test.go:355
Error: Not equal:
expected: "stub"
actual : ""
Diff:
--- Expected
+++ Actual
@@ -1 +1 @@
-stub
+
Test: TestDial_Tracing/stream
--- FAIL: TestDial_Tracing (0.02s)
--- PASS: TestDial_Tracing/unary (0.01s)
--- FAIL: TestDial_Tracing/stream (0.01s)
In an attempt to fix these flaky tests, ensure Finish() is called on all instances of opentracing.Span.
[1] #3402 (closed)
Edited by Toon Claes