[7u] problem with jdk test java/nio/channels/DatagramChannel/SendToUnresolved.java

Francis ANDRE francis.andre.kampbell at orange.fr
Mon Oct 28 10:34:14 PDT 2013


Hi Alan

Yes, This is the second time I saw an resolved address for an non resolvable 
host. There is only 2 PC connected locally that are going to internet thru the 
ISP Orange. So I am a little bit puzzled with those tests

Le 28/10/2013 08:57, Alan Bateman a écrit :
> On 28/10/2013 04:10, Francis ANDRE wrote:
>> Hi
>>
>> On my WXP/Cygwin/VS2010 platform, this jdk test 
>> java/nio/channels/DatagramChannel/SendToUnresolved.java fails because the
>> TestUtil.UNRESOLVABLE_HOST which is the string "blah-blah.blah-blah.blah" 
>> gets resolved to blah-blah.blah-blah.blah/66.152.109.63:37
>>
>> So, I am wondering why blah-blah.blah-blah.blah gets resolved. What should be 
>> the corrective action: ignoring this test?
>>
>> Francis
> In your other mail then you proposed a patch to 
> test/java/net/URL/OpenStream.java as it is resolving foo.bar.baz when the test 
> doesn't expect it. So it's hard to say, it suggests maybe something in the 
> Windows configuration. Do you know if all unknown hosts are resolving to this 
> address on this system?
Yes, all ping requests to a host name composed by at least 2 subnames separated 
by a '.' are localized to [66.152.109.63]. Only single host name, ie without 
subdomain, are not localized. Ok, I found the problem which is coming from my 
DynamicDSN provider. Removing it's own DSN addresses from the DSN list makes 
blah-blah.blah-blah.blah unresolved.

So I am sorry for the noise caused by those patches which are wrong for this 2 
tests: SendToUnresolved.java and OpenStream.java

Francis

>
> -Alan
>




More information about the jdk7u-dev mailing list