<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi Alan,<br>
<br>
Since this patch is for Unix like platforms, I didn't test the test
case on Windows yet.<br>
<br>
However when I tested it on Windows I found it will fail on Windows.<br>
<br>
I also use ping command on Windows to test both 0.0.0.0 and ::0,
they all fail. So I want to know if we need further investigation to
see why these 2 addresses can not be reachable or we just think it
is the proper behavior because JDK's behavior is consistent with
ping on Windows?<br>
<br>
Thanks a lot!<br>
<br>
On 05/07/2012 03:46 PM, Alan Bateman wrote:
<blockquote cite="mid:4FA77DC0.5020901@oracle.com" type="cite">
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
On 07/05/2012 03:29, Deven You wrote:
<blockquote cite="mid:4FA7337B.1020901@linux.vnet.ibm.com"
type="cite">
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
Hi Alan and Chris,<br>
<br>
I have updated the webrev[1] according to your suggestions.<br>
<br>
For the test case, I did two things:<br>
<br>
1. add the @run main/othervm -Djava.net.preferIPv4Stack=true
PingThis<br>
<br>
2. Only when !preferIPv4Stack and hasIPv6(), ::0 will be added
into the address list.<br>
<br>
I think it's enough for the test case, please review it.<br>
<br>
I just have another concern, how we can configure the test env
so this test case can be run under root (may use sudo)
privilege?<br>
<br>
[1]
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
<a moz-do-not-send="true"
href="http://cr.openjdk.java.net/%7Elittlee/OJDK-217/webrev.04/">http://cr.openjdk.java.net/~littlee/OJDK-217/webrev.04/</a><br>
</blockquote>
I think we are close to the finish line on this one.<br>
<br>
One thing I notice in the webrev is that net_util_md.h hasn't been
updated to defined the prototype for
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
<span class="new">NET_IsZeroAddr. It would be good to add that (no
need to generate another webrev just for this one).<br>
<br>
The test case looks okay to me. Minor comment is that the while
loop could be replaced with for (String: addr: addrs) { ...}. No
need to re-generate the webrev if you decide to take up this
suggestion.<br>
<br>
I don't think we should change the test to attempt to run it via
sudo, that would just complicate the test as there is no
guarantee that the user running the tests is in sudoers. Also
one wouldn't want the test prompting for a password. it would
also require a script as it would need to be skipped on
platforms such as Windows.<br>
<br>
So I think I'm happy with the test as is and I assume you will
run with regular user/root before pushing. I also assume you
will run it on Windows to make sure that it passes there too.<br>
<br>
-Alan.<br>
</span> </blockquote>
<br>
<br>
<pre class="moz-signature" cols="72">--
Best Regards,
Deven</pre>
</body>
</html>