DatagramPacketTest2

DatagramPacketTest2

[Dalibor Topic]

Hello mauve hackers,

I think the zero buffer length invalid receive data
test from
gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2
is not correct. The test expects to see an
IOException, while the (1.4) spec says that receive is
supposed to block until something is received, and
truncate it if necessary.

I came accross it since it's the only test that just
plain hangs when I run it in both latest kaffe from
CVS, as well as with Sun's JDK 1.3.

best regards,

dalibor topic


__________________________________________________
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com

Re: DatagramPacketTest2

[John Leuner]

> I think the zero buffer length invalid receive data
> test from
> gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2
> is not correct. The test expects to see an
> IOException, while the (1.4) spec says that receive is
> supposed to block until something is received, and
> truncate it if necessary.
> 
> I came accross it since it's the only test that just
> plain hangs when I run it in both latest kaffe from
> CVS, as well as with Sun's JDK 1.3.

Have a fix?

John

Re: DatagramPacketTest2

[Mark Wielaard]

Hi,

On Tue, 2002-04-30 at 19:21, Dalibor Topic wrote:
> I think the zero buffer length invalid receive data
> test from
> gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2
> is not correct. The test expects to see an
> IOException, while the (1.4) spec says that receive is
> supposed to block until something is received, and
> truncate it if necessary.
> 
> I came accross it since it's the only test that just
> plain hangs when I run it in both latest kaffe from
> CVS, as well as with Sun's JDK 1.3.

We have it disabled in the mauve configure file from libgcj for that
very same reason. I think you should feel free to remove that check from
the test.

Cheers,

Mark

Re: DatagramPacketTest2

[Dalibor Topic]

Hi John,

--- John Leuner <jewel@pixie.co.za> wrote:
> > I think the zero buffer length invalid receive
> data
> > test from
> >
>
gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2
> > is not correct. The test expects to see an
> 
> Have a fix?

I'll post a patch that removes the test.

best regards,

dalibor topic

__________________________________________________
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com

Re: DatagramPacketTest2

[Dalibor Topic]

[-- Attachment #1: Type: text/plain, Size: 419 bytes --]

A patch removing the test is attached.

changelog:
 *
gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2.java:
deleted broken zero lenth buffer test.

--- Dalibor Topic <robilad@yahoo.com> wrote:
> 
> I'll post a patch that removes the test.

happy patching,

dalibor topic

__________________________________________________
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com

[-- Attachment #2: mauve.diff --]
[-- Type: application/octet-stream, Size: 691 bytes --]

diff -ur mauve/gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2.java new-mauve/gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2.java
--- mauve/gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2.java	Tue Apr  2 20:23:32 2002
+++ new-mauve/gnu/testlet/java/net/DatagramSocket/DatagramSocketTest2.java	Wed May  1 02:49:39 2002
@@ -229,17 +229,6 @@
 			harness.check(false, "Expected IOException");
 		}
 
-		try
-		{
-			// zero data buffer length
-			DatagramPacket pkt = new DatagramPacket(buf, 0);
-			sock.receive(pkt);
-			harness.check(true);
-		}
-		catch (IOException e)
-		{
-			errormsg("invalid_receive_data", 5, false, "IOException");
-		}
 		sock.close();
 	}