activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David Keaveny (JIRA)" <jira+amq...@apache.org>
Subject [jira] Issue Comment Edited: (AMQNET-127) Add support for Sessions
Date Mon, 08 Dec 2008 05:46:05 GMT

    [ https://issues.apache.org/activemq/browse/AMQNET-127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=47921#action_47921
] 

ctrlaltdel edited comment on AMQNET-127 at 12/7/08 9:45 PM:
---------------------------------------------------------------

20080808-DK-AddSupportForOneWaySessions.patch (attached) is a preliminary implementation of
one-way (datagram) session support.

I would appreciate any suggestions people have on how best to automate testing for all this.
I created a test project which runs a ServiceHost internally and calls from a client in the
same project, and for a few tests, this works OK; however, as the number of tests increase,
the results become increasingly non-deterministic, to the point where they become meaningless.


Testing the one-way code is particularly difficult, as WCF's implementation means you can't
easily go peeking at the contents of the service. I tried serialising the service's state
to disk and reading that in from the tests, but came across all sorts of synchronisation errors,
where I was was reading the file before it was being written to. Adding random Thread.Sleeps
feels very wrong, and of course makes the tests even less scaleable.

Again, if anyone has ideas on how to automate testing the WCF provider itself, I would be
very interested to hear them.

      was (Author: ctrlaltdel):
    Preliminary implementation of one-way (datagram) session support.

I would appreciate any suggestions people have on how best to automate testing for all this.
I created a test project which runs a ServiceHost internally and calls from a client in the
same project, and for a few tests, this works OK; however, as the number of tests increase,
the results become increasingly non-deterministic, to the point where they become meaningless.


Testing the one-way code is particularly difficult, as WCF's implementation means you can't
easily go peeking at the contents of the service. I tried serialising the service's state
to disk and reading that in from the tests, but came across all sorts of synchronisation errors,
where I was was reading the file before it was being written to. Adding random Thread.Sleeps
feels very wrong, and of course makes the tests even less scaleable.

Again, if anyone has ideas on how to automate testing the WCF provider itself, I would be
very interested to hear them.
  
> Add support for Sessions
> ------------------------
>
>                 Key: AMQNET-127
>                 URL: https://issues.apache.org/activemq/browse/AMQNET-127
>             Project: ActiveMQ .Net
>          Issue Type: New Feature
>          Components: WCF
>    Affects Versions: 1.1
>            Reporter: Jim Gomes
>            Assignee: Jim Gomes
>             Fix For: 1.1
>
>         Attachments: 20080808-DK-AddSupportForOneWaySessions.patch
>
>
> Add support for Sessions in WCF provider.  Following is a link to a blog that gives an
overview of how sessions work in WCF.
> [http://www.pluralsight.com/community/blogs/aaron/archive/2006/02/27/19253.aspx]

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message