Kaffe built statically on SunOS4.1.3_U1 - some questions

Joe Hartley jh at brainiac.com
Sun Jan 3 14:56:24 PST 1999

First, thanks to Godmar for pointing me to the FAQ regarding SunOS 4.1.3
dynamic library problems.  I was able to compile statically fine when I
made the substitution of bcopy() for memmove().  I've seen that so many
times before it's now a trivial thing for me to work around!

Now that I've compiled, I have a couple of questions.  I couldn't find an
answer easily in the archives, but they're not searchable on keywords, so
I can't tell if I'm really rehashing known territory.

First, I get as far as the UDPTest when I run "gmake test", then everything
just seems to freeze.  Is there a known problem with this test under SunOS?
I see in a ps that the test process is still there, but it's not doing
 2966 p1 IW    0:00 gmake test
 2968 p1 IW    0:00 /bin/sh -c (cd test && gmake test)
 2969 p1 IW    0:00 gmake test
 2972 p1 IW    0:00 /bin/sh -c
 2974 p1 IW    0:00 /bin/sh ./TestScript
 3206 p1 IW    0:02 /usr/local/bin/Kaffe UDPTest

I removed the UDPtest from TestScript in the test/regression directory,
and it passed all tests.  Only UDPTest gives any problem.

My other question may be related to the first.  Some background:  I am
trying to run some Java tools provided by Lucent to manage a Portmaster.
When I try to run them, they complain that the login to the Portmaster
has not succeeded, though there is output in a log file that can only have
been received after a successful login.

I've included below the debug output I got and some comments from a
tech at Lucent.

>        at java/io/InputStreamReader.read(52)
>        at java/io/BufferedReader.refillBuffer(163)
>        at java/io/BufferedReader.readLine(129)
>        at COM/livingston/model/LEDataSpecFactory.createFrom(70)
>        at COM/livingston/model/LEDataSpecFactory.GetPortMasterSpec(53)
>        at COM/livingston/model/LEPortMaster.<init>(140)
>        at CLUtil.connectToPortMaster(104)
>        at PMCommand.connectToPortMaster(363)
>        at PMCommand.interactiveCommands(304)
>        at PMCommand.processArgs(150)
>        at PMCommand.main(44)
>Exception handled at:
>        at java/lang/Throwable.<init>(31)
>        at java/lang/Exception.<init>(17)
>        at COM/livingston/util/LEDebugLog.dumpStack(181)
>        at COM/livingston/util/LEDebugLog.logln(54)
>        at CLUtil.connectToPortMaster(111)
>        at PMCommand.connectToPortMaster(363)
>        at PMCommand.interactiveCommands(304)
>        at PMCommand.processArgs(150)
>        at PMCommand.main(44)
>This bugs me. It looks like you've successfully connected and were trying to
>read a file in the pmtools.jar file which contains some configuration
>information.  The BufferedReader (java/io/BufferedReader.readLine)
>apparently could read the file and decided to throw a null pointer
>exception.  My code receives this NullPointerException and assumes that the
>connection could not be made. 
>Do you know which Java Virtual Machine Kaffe is supposed to emulate?  This
>definitely looks like a Kaffe problem.  You might want to check and make
>sure that your Kaffe is the latest version.  If possible, make sure they
>emulate at least Java 1.1.4, preferably 1.1.6 or later.

I am not familiar enough with the Java specs to know how PersonalJava 1.1
(which is Kaffe's target emulation) compares with the numbers being used
by the Lucent tech.

It has occured to me that there may be some networking problem with Kaffe
on my system.  The Lucent tools do not use UDP, but do use TCP on port 1643.
If there _is_ a networking problem that causes the UDPTest to fail, it could
be directly related to the problems I'm having with the Lucent tool.

Thanks in advance for any clues or pointers!

        Joe Hartley - jh at brainiac.com - brainiac services, inc
 PO Box 5069 : Greene, RI : 02827 - vox 401.539.9050 : fax 401.539.2070
  Without deviation from the norm, "progress" is not possible. - FZappa

More information about the kaffe mailing list