<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Alexis Richardson wrote:
<blockquote
cite="mid:167204d20901050855l770443aay745999567913035f@mail.gmail.com"
type="cite">
<pre wrap="">Carl,
On Mon, Jan 5, 2009 at 4:17 PM, Carl Trieloff <a class="moz-txt-link-rfc2396E" href="mailto:cctrieloff@redhat.com"><cctrieloff@redhat.com></a> wrote:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">If queue lifecyle were defined and its operations were atomic, then
that would make it is easier for TX to cover all atomicity cases.
</pre>
</blockquote>
<pre wrap="">... so a queue that is created
during a TXN flow would outlive
a TX abort as it is not a transnational resource.
</pre>
</blockquote>
<pre wrap=""><!---->
Given that you are saying that a queue is not a transactional resource
here, </pre>
</blockquote>
<br>
let me be specific, AMQP today does not provide a way to enlist a
queue's creation or deletion into<br>
a txn context. This in no way affects the ability to provide ACID
transactional gaurentees on messages.<br>
<br>
It does however imply some implementation details that need to be taken
care of.<br>
Carl.<br>
</body>
</html>