The Jira is the way to track this. Since you own the issue, people will
assume that you own it. Woe to the person who starts working on it
w/out owning the Jira issue.
BTW, thanks for picking this up!
Regards,
Alan
Mosur Ravi, Balaji wrote:
I had started on it but have been side tracked to fixing other issues.
Please let me know if anyone is picking this up...
- Balaji
-----Original Message-----
From: Kulp, John Daniel
Sent: Friday, June 02, 2006 4:03 PM
To: [email protected]
Cc: Alan D. Cabrera; [EMAIL PROTECTED]
Subject: Re: Test failure and hang building latest code
Balaji started working a little bit on this this morning and ran into
some
snags. I'm not sure how far he got, but it MAY not be possible with
the way the Maven surefire plugin works. If you look at the code for
java.rmi.RMIClassLoader, the initialize method does:
Class providerClass =
Class.forName(providerClassName, false,
ClassLoader.getSystemClassLoader());
Thus, the class has to be on the SYSTEM classpath, it doesn't use any
type
of context classloader. When maven "forks" surefire, it sets a system
classpath that ONLY contains the Surefire stuff which then loads the
project classpath in an internal classloader.
Anyway, the code probably works, but I'm not sure how to get surefire to
actually use it. We MAY need to put just that class in a separate jar
and maybe add a "<argLine>-Xbootclasspath/p:thatjar</argLine>" section
to the surefire config. (requires forking be turned on) I'm not
really sure yet.
Dan
On Friday 02 June 2006 15:43, Alan D. Cabrera wrote:
This is due to the fact that the file path to your local maven
repository has spaces in it, "Documents and Settings". The default
RMI
classloader barfs. We should use the RMI classloader that's in
Geronimo,
modules/system/src/java/org/apache/geronimo/system/rmi/RMIClassLoaderSp
iImpl.java
It's seems to be suitably licensed for us to poach. :)
Regards,
Alan
Mosur Ravi, Balaji wrote:
Hi,
This was mentioned earlier & I think there is a JIRA raised for
this...
- Balaji
-----Original Message-----
From: Rick McGuire [mailto:[EMAIL PROTECTED]
Sent: Friday, June 02, 2006 6:38 AM
To: [email protected]
Subject: Test failure and hang building latest code
I'm getting a test failure trying to build the latest code level (a
fresh checkout). This exception is reported, and the build just
hangs at this point. Is anybody else having this problem?
Rick
-------------------------------------------------------
T E S T S
-------------------------------------------------------
[client:err] Exception in thread "main" java.rmi.UnmarshalException:
error unmar
shalling return; nested exception is:
[client:err] java.net.MalformedURLException: no protocol: and
[client:err] at sun.rmi.registry.RegistryImpl_Stub.lookup(Unknown
Source)
[client:err] at
org.apache.yoko.processmanager.internal.ProcessAgentImpl.main
(ProcessAgentImpl.java:65)
[client:err] Caused by: java.net.MalformedURLException: no protocol:
and [client:err] at java.net.URL.<init>(URL.java:567)
[client:err] at java.net.URL.<init>(URL.java:464)
[client:err] at java.net.URL.<init>(URL.java:413)
[client:err] at
sun.rmi.server.LoaderHandler.pathToURLs(LoaderHandler.java:74
7)
[client:err] at
sun.rmi.server.LoaderHandler.loadClass(LoaderHandler.java:147
)
[client:err] at
java.rmi.server.RMIClassLoader$2.loadClass(RMIClassLoader.jav
a:620)
[client:err] at
java.rmi.server.RMIClassLoader.loadClass(RMIClassLoader.java:
247)
[client:err] at
sun.rmi.server.MarshalInputStream.resolveClass(MarshalInputSt
ream.java:197)
[client:err] at
java.io.ObjectInputStream.readNonProxyDesc(ObjectInputStream.
java:1538)
[client:err] at
java.io.ObjectInputStream.readClassDesc(ObjectInputStream.jav
a:1460)
[client:err] at
java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStrea
m.java:1693)
[client:err] at
java.io.ObjectInputStream.readObject0(ObjectInputStream.java:
1299)
[client:err] at
java.io.ObjectInputStream.readObject(ObjectInputStream.java:3
39)
[client:err] ... 2 more
