After checking out the cvs head(08-30),I run the following commands: mkdir $BUILD ../frysk-0830/autogen.sh; make -j2 All is ok till now. Then launch frysk through "frysk-gui/fryk/gui/FryskGui" and I got the following information: [root@YZ_X86_64 /rtos/source/frysk/temp/build_0830]# frysk-gui/frysk/gui/FryskGui GTK Accessibility Module initialized Unable to get host name or ip address from getHostAddress()/getHostName() java.net.UnknownHostException: YZ_X86_64 at java.net.InetAddress.getLocalHost(libgcj.so.7) at java.net.InetAddress.getLocalHost(libgcj.so.7) at frysk.gui.common.IconManager.getHostName(FryskGui) at frysk.gui.common.IconManager.<clinit>(FryskGui) at java.lang.Class.initializeClass(libgcj.so.7) at frysk.gui.common.IconManager.setImageDir(FryskGui) at frysk.gui.Gui.gui(FryskGui) at frysk.gui.FryskGui.main(FryskGui) Caused by: java.net.UnknownHostException: YZ_X86_64 at java.net.InetAddress.lookup(libgcj.so.7) at java.net.InetAddress.getLocalHost(libgcj.so.7) ...7 more Exception in thread "main" java.lang.ExceptionInInitializerError at java.lang.Class.initializeClass(libgcj.so.7) at frysk.gui.common.IconManager.setImageDir(FryskGui) at frysk.gui.Gui.gui(FryskGui) at frysk.gui.FryskGui.main(FryskGui) Caused by: java.lang.NullPointerException at java.net.InetAddress.getHostAddress(libgcj.so.7) at frysk.gui.common.IconManager.getHostName(FryskGui) at frysk.gui.common.IconManager.<clinit>(FryskGui) at java.lang.Class.initializeClass(libgcj.so.7) ...3 more In PPC64 and X86, there's no any exceptions. We thought the history log files may cause this bug, but it doesnot work after we removed all logs under /root/.frysk*.
I think this might be a simple Java bug, does a simple call to java.net.InetAddress.getLocalHost return anything sane?
Bizarre, I tested this on both x86 and x86_64 and did not have this problem. Hmmmm...
I cannot seem to duplicate this bug on my x86 or x86_64 machines this mornig with a fresh checkout/build. Here are the versions of the kernel/compiler/os i am running on those machines: x86: uname -a = "Linux localhost.localdomain 2.6.17-1.2174_FC5 #1 Tue Aug 8 15:30:55 EDT 2006 i686 i686 i386 GNU/Linux" gcj --version = "gcj (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)" os: FC5 fully updated x86_64: uname -a = "Linux frysk.boston.redhat.com 2.6.17-1.2586.fc6 #1 SMP Thu Aug 24 08:51:26 EDT 2006 x86_64 x86_64 x86_64 GNU/Linux" gcj --version = "gcj (GCC) 4.1.1 20060825 (Red Hat 4.1.1-19)" os: FC6 test 2 Hmmmm, like Andrew said, wonder if we are seeing some small java bug here. Guess I can add something that is that exception is thrown to put some sort of error in the tootip or null out all fields for the tooltip. uname -a = "
Lets get the bug fixed; I suspect a misconfigured host. A testcase should be simple, just call getLocalHost.
Added a test in TestIconManager.java where it checks for the validity of the call to getLocalHost. A "make check" in frysk-gui will run the test. I have also modded IconManager.java to not stacktrace when it cannot get valid host information. I'm not sure a misconfured host is necessarily a valid reason to stop frysk from coming up. Since I cannot duplicate the failure on the systems I have available Yong, please verify these changes.
> > x86_64: > uname -a = "Linux frysk.boston.redhat.com 2.6.17-1.2586.fc6 #1 SMP Thu Aug 24 > 08:51:26 EDT 2006 x86_64 x86_64 x86_64 GNU/Linux" > > gcj --version = "gcj (GCC) 4.1.1 20060825 (Red Hat 4.1.1-19)" > > os: FC6 test 2 > Versions of our X86_64 : uname -a = Linux YZ_X86_64 2.6.15-1.2054_FC5 #1 SMP Tue Mar 14 15:48:20 EST 2006 x86_64 x86_64 x86_64 GNU/Linux gcj --version = "gcc (GCC) 4.1.1 20060525 (Red Hat 4.1.1-1)" I will update our gcj packages and try it again. It should be one bug in java packages.
Did updating resolve your problem?
(In reply to comment #7) > Did updating resolve your problem? I think there are some problems on my local java packages(such as: libgcj). I wrote one small java program which just calls the java.net.NetAddress.getHostAddress(), but I still got same exceptions after finishing compiling it. I want to update the java package but failed on FC5, So I have to wait for FC6 release version. Ah, the bug is not the problem of your codes(it's just my local OS's problem.). :-):-)
Apparently fixed. Cannot reproduce in current checkout from HEAD.