youngoli commented on a change in pull request #11413: [BEAM-9746] check for 0
length copies from state
URL: https://github.com/apache/beam/pull/11413#discussion_r408481881
##########
File path: sdks/go/pkg/beam/core/runtime/harness/statemgr_test.go
##########
@@ -258,6 +261,167 @@ func TestStateChannel(t *testing.T) {
}
}
+// TestStateKeyReader validates ordinary Read cases
+func TestStateKeyReader(t *testing.T) {
+ const readLen = 4
+ tests := []struct {
+ name string
+ buflens []int // sizes of the buffers received on the state
channel.
+ numReads int
Review comment:
Is the number of reads something that needs to be predictable to whatever
code uses the reader? To me it seems like one of those implementation details
that doesn't need to be unit tested because it's invisible to the caller.
The reason this stood out to me is that the `numReads` in all the test cases
below seem a bit inscrutable, and it would cause these tests to fail if the
implementation of the `stateKeyReader` was changed slightly, which is just an
annoyance unless it would actually break code.
On the other hand, if the number of reads _is_ a detail that users should
know, and correctness does rely on it staying consistent, that seems like it
should be explicitly documented on `stateKeyReader.Read`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
