[jcifs] jcifs bug

[Michael B.Allen]

On Fri, 22 Mar 2002 10:30:15 -0700
"Bernie Wieser" <bernie.wieser at calgary.qcdata.com> wrote:

> I will try this fix, as it takes in to account the security authinfo I
> ignored.
> 
> I noticed you are now relying on the default urlstreamhandler to do
> toExternalForm.  Excellent;
> 
> The only odd issue we have is if I compile the .61 code with 131_02,
> we get a thread deadlock issue.

So does this have anything to do with the handler? Make sure you check
to see if your getting dead air from the initial SYN packet. That takes
around 3 min 15 sec to timeout with "No route to host".

>  I have yet to figure this out.
> I had to resort to only changing one file in the .60 release jar,
> as the release .61 jar hangs on our linux environment.

Which file would that be?

>  Its funny;
> we have the same configuration on development (w2k), testing (linux),
> and production (solaris).  Yet one thing or another fails to
> work in our software when we redeploy.  (This is not re jcifs specifically.)
> 
> Thanks for getting back to me.
> 
> B.
> 
> p.s. the test case  ...
> URL urls[] = new URL[10];
> for( int i = 0; i < 10; i++ )
> urls[i] = new URL( "smb://somehost/someshare/"+i );
> for( int i = 0; i < 10; i++ ){
> System.out.println( urls[i].toString() );
> if( i > 0 ) if( urls[i].toString().equals( urls[i-1].toString() ) )
> fail("miserably");

This hangs? That's pretty strange because it's not even threaded. See
the examples directory for threaded tests that might be better adapted
to your URL handler unit tests. You might also want to look very closely
at DNS. DNS problems can cause all sorts of wierd hanging-like scenarios.

Good luck,
Mike

-- 
May The Source be with you.