Halts during the NioWrapperTest

[GoogleCodeExporter]

I am installing bitcoinj using Maven. mvn install. And the install halts during 
the NioWrapperTest. See attached screenshot. I skip the test (mvn 
-Dmaven.test.skip=true install) and am able to execute code.

Original issue reported on code.google.com by jad...@adairservices.net on 25 Sep 2013 at 8:50

Attachments:

• HaltedDuringMavenInstallScreenShot.jpg

[GoogleCodeExporter]

Is this reliably reproducible, or a race?

Original comment by BlueMatt...@gmail.com on 25 Sep 2013 at 8:53

[GoogleCodeExporter]

You'll have to excuse me: I do not know what you mean by it being "a race". It 
is reproducible.

Original comment by jad...@adairservices.net on 25 Sep 2013 at 8:57

[GoogleCodeExporter]

Are you running some kind of local firewall?

This might be a Windows-specific issue too. AFAIK none of us develop on 
Windows. Of course that's not meant to happen in Java, but occasionally things 
leak through the abstractions.

Original comment by hearn@google.com on 30 Sep 2013 at 8:21

[GoogleCodeExporter]

Only Windows Firewall which has exceptions in it to allow bitcoind, bitcoin-qt, 
and armory through... but disabling the firewall entirely still yields the same 
result (i.e., halting during the NioWrapperTest).

Oddly, I read somewhere that the M2_HOME variable is not needed anymore for 
mvn-3.x; So I deleted it and the mvn install still went to completion, but with 
errors. I have attached that screenshot and some output surefire-reports.

If I was a betting man, I would bet with you in the hunch that it is a 
Windows-specific issue. FYI: Apparently Maven 3.1 is using SLF4J as the core 
for logging.

Original comment by jad...@adairservices.net on 30 Sep 2013 at 4:57

Attachments:

• com.google.bitcoin.protocols.niowrapper.NioWrapperTest.txt

[GoogleCodeExporter]

Um, attached file.

Original comment by jad...@adairservices.net on 30 Sep 2013 at 4:59

Attachments:

• NioWrapperTestMavenInstallScreenShot.jpg

[GoogleCodeExporter]

Hrmmm. Could you run the tests in an IDE and see if it reproduces there?

That test listens on a local socket and then connects back to it. The error 
means it can't connect back to itself for some reason. I've not seen such a 
failure let alone a reproducible one.

Could you step through it with a debugger and see if the local socket is 
actually open and whether you  can connect to it with telnet?

Original comment by hearn@google.com on 30 Sep 2013 at 5:13

[GoogleCodeExporter]

Closing this due to the nonetty merge. If you can reproduce this on git master 
please reopen.

Original comment by hearn@google.com on 12 Dec 2013 at 3:11

• Changed state: Fixed

[GoogleCodeExporter]

I had the same problem. Firewall was the reason...

Original comment by Andrej.M...@gmail.com on 21 Jan 2014 at 4:41