[
https://issues.apache.org/jira/browse/DISPATCH-1820?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17223908#comment-17223908
]
Charles E. Rolke commented on DISPATCH-1820:
--------------------------------------------
As-is the test hangs about half of the time.
Adding a short sleep() to the echo client gets the test to pass hundreds of
times in a row.
{code:java}
iff --git a/tests/TCP_echo_client.py b/tests/TCP_echo_client.py
index 57152e3d..80972451 100755
--- a/tests/TCP_echo_client.py
+++ b/tests/TCP_echo_client.py
@@ -144,6 +144,7 @@ class TcpEchoClient():
selectors.EVENT_READ | selectors.EVENT_WRITE)
# event loop
+ time.sleep(0.1)
while self.keep_running:
if self.timeout > 0.0:
elapsed = time.time() - start_time
{code}
> TCP adaptor test time out with single one-byte message
> ------------------------------------------------------
>
> Key: DISPATCH-1820
> URL: https://issues.apache.org/jira/browse/DISPATCH-1820
> Project: Qpid Dispatch
> Issue Type: Bug
> Components: Protocol Adaptors
> Reporter: Charles E. Rolke
> Priority: Major
> Attachments: DISPATCH-1820-log-ok.txt, DISPATCH-1820-log-stall.txt
>
>
> Running the simplest system_tests_tcp_adaptor (one router, one one-byte
> message) sometimes works and sometimes times out. Combing through the router
> logs for the ok and stall cases reveals some differences in adapter/core
> events.
> These logs are for a test that sends a one byte message and a 100 byte
> message. In the stall case the first one-byte message never goes through.
> The logs have been sanitized for easier diffing:
> * tcp ports reanamed from numbers to 'listener' and 'connector'
> * time-of-day deleted
> * delivery incref and decref lines deleted
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
