Re: [sci-flex-cvs] [sci-flex commit] r367 - wiki

0 views
Skip to first unread message

Senaka Fernando

unread,
Dec 12, 2008, 11:32:10 PM12/12/08
to sci-fl...@googlegroups.com
Harsha,

Aren't we going to test samples with the automation tests?

Regards,
Senaka

On Sat, Dec 13, 2008 at 7:02 AM, <codesite...@google.com> wrote:

Author: Harsha.Halgaswatta
Date: Fri Dec 12 17:11:33 2008
New Revision: 367

Modified:
   wiki/MediatorToDo.wiki

Log:
Edited wiki page through web user interface.

Modified: wiki/MediatorToDo.wiki
==============================================================================
--- wiki/MediatorToDo.wiki      (original)
+++ wiki/MediatorToDo.wiki      Fri Dec 12 17:11:33 2008
@@ -7,6 +7,6 @@
   # ~~JMS Map Messages should be converted 2 a Map Backed OM, ONLY IF
NEEDED. Use a Wrapped representation~~
   # ~~Sample(s) on Synapse that uses the JMS !MapMessage implementation~~
   # write sci-flex samples for Map scenarios, and epl query based filter.
*Note:* This is only for Axiom Mediator.
-  # Test each and every samples which illustrates different scenarios in
SCI-Flex using !AbstractAutomationTestCase
+  # ~~Test each and every samples which illustrates different scenarios in
SCI-Flex using !AbstractAutomationTestCase~~
   # Test every posible scenarios for newly added underlying logics
   # Application of Concurrent Programming principals



Harsha Halgaswatta

unread,
Dec 14, 2008, 2:46:00 AM12/14/08
to sci-fl...@googlegroups.com
Hi Senaka,

Synapse folks have used automation tests as they are having many samples those are functioning differently in different areas. But in our case we have tested our main objective which is the connection between Synapse and Esper using XML and Axiom Mediators even with the Registry support to handle EPL statements and End points, using 25 test scenarios. Hence my opinion is we don't need to do automation tests in SCI-Flex.


Cheers
Harsha
Reply all
Reply to author
Forward
0 new messages