Nabble - JamVM

Re: MIPS

2009-11-24T17:47:11Z

Hi Michel,

2009/11/16 michel-timos <michel@...>:

>
> Hello,
>
> I am currently looking to port the 'inline' part of the latest 1.5.3 version
> to the MIPS platform so we can benefit from the performance increase
> inlining brings.
>
> Does anyone have any information on how to do that? Or where to look for
> documentation on how this inlining part works?
>
> It does not need to be a ready made solution (although I wouldn't mind :-)
> ), but just a tip or hint would already help me a lot.
>

I didn't implement the code-copying JIT on MIPS because I don't
currently have a MIPS machine to develop or test on (I've been meaning
to investigate QEMU but haven't had time). There isn't a lot of
platform specific bits needed, just two macros defined in arch/mips.h
:

FLUSH_CACHE(address, length)

This is required to flush the data and instruction caches so that
generated machine instructions can be executed.

GEN_REL_JMP(target_addr, patch_addr, patch_size)

This is used to generate a relative jump. The jump target address is
target_addr, and the address where the instruction should be generated
is patch_addr. The patch_size is the amount of room available at
patch_addr.

GEN_REL_JMP is used as an expression (it's mostly implemented as a
statement expression). The result is true or false, depending on
whether the jump could be generated. For example, the result will be
false if the target is out of range of the jump instruction, or there
isn't enough room to generate the instruction. This means GEN_REL_JMP
can be minimally implemented as:

#define GEN_REL_JMP(target_addr, patch_addr, patch_size) false

This will impact performance, but not by much.

Hope this helps,
Rob.

> Thanks
> --
> View this message in context: http://old.nabble.com/MIPS-tp26370456p26370456.html
> Sent from the JamVM mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Jamvm Performance

2009-11-23T11:28:36Z

Robert Lougher wrote:

> Adding -verbosegc to the command line will show when a gc occurs, and
> how long it takes.
>
> Rob
>
> 2009/11/12 Arigead <captain.deadly@...>:
>> David Griffiths wrote:
>>> I believe JamVM uses a "stop-the-world" Garbage Collector, so called
>>> because when you run out of heap it has to stop all threads while the
>>> Garbage Collector (GC) runs. The GC pause can be quite long depending
>>> on how many objects are in the heap. There are other JVMs out there
>>> that minimise this pause through the use of more sophisticated GC
>>> algorithms but at the cost of increased complexity and footprint.
>>>
>>> Cheers,
>>>
>>> Dave
>>>
>>> On Thu, Nov 12, 2009 at 3:01 PM, Arigead <captain.deadly@...> wrote:
>>>> Hi David,
>>>> Sorry about this, new to Java, well this indept part of java but
>>>> what is the GC Cycle?
>> I should have asked in that previous email is there any way that I can
>> enable some debug in the Jamvm build I have to see if this is the
>> problem. I'm building from source so if there's debug I can easily
>> switch it on if that'll shed some light onto things. Of course the
>> overhead of debug might not help me.
>>
>> Thanks for your help
>>

Hello Me again,
thanks a million for all your previous help on this but I fear I may
be on the wrong list but I'm sure it's going to be a very difficult
problem to pin down. I've looked into running jamvm with the -vorbosegc
switch and I'll paste a little run of my OSGi Bundle at the end of this
email.

The reason I say that this is going to be hard to pin down is that I now
have my code as both a simple stand alone Java App and a Simple OSGi
Java Bundle. All the code does is establish a DBus connection and
"register an interest" in the RhythmBox music player status. I'm running
on linux, if that's an issue.

So under Sun Java 6 both the Stand alone App and the OSGi Bundle running
in Knophleurfish run perfect.

Under Jamvm the Stand alone App runs perfect but the OSGi Bindle are not
getting events. Perhaps I should be talking to an OSGi List but I'm sure
they'll just say that if the Bundle runs perfect under Sun Java 6 then
it's not their problem.

Likewise the members of this list would be well with in their rights to
say if the Stand Alone App works then it's obviously an OSGi Problem and
not a Jamvm Problem.

Oh well a cup of tea and a ponder ;-) Anyhow the verbose output gives me
a regular stream of garbage collection but not what I'd call excessive.
Perhaps I could try all this on my bigger 64Bit faster Server then this
laptop, which I like to work on.

If anybody has any more ideas I'd be only too happy to hear them

Thanks

<GC: Alloc attempt for 576 bytes failed.>
<GC: Allocated objects: 110699>
<GC: Freed 41277 object(s) using 6692040 bytes>
<GC: Largest block is 115384 total free is 9927264 out of 16777208 (59%)>
<GC: Mark took 0.132633 seconds, scan took 0.003852 seconds>
<GC: running 87 finalisers>
<GC: Alloc attempt for 32784 bytes failed.>
<GC: Allocated objects: 110252>
<GC: Freed 39070 object(s) using 6580352 bytes>
<GC: Largest block is 156240 total free is 10077952 out of 16777208 (60%)>
<GC: Mark took 0.104187 seconds, scan took 0.003806 seconds>
<GC: running 85 finalisers>
<GC: Alloc attempt for 7472 bytes failed.>
<GC: Allocated objects: 109862>
<GC: Freed 32370 object(s) using 6740600 bytes>
<GC: Largest block is 122016 total free is 10116952 out of 16777208 (60%)>
<GC: Mark took 0.085880 seconds, scan took 0.003669 seconds>
<GC: running 64 finalisers>
<GC: Alloc attempt for 32784 bytes failed.>
<GC: Allocated objects: 111505>
<GC: Freed 38972 object(s) using 6719072 bytes>
<GC: Largest block is 156240 total free is 9935288 out of 16777208 (59%)>
<GC: Mark took 0.112957 seconds, scan took 0.003821 seconds>
<GC: running 85 finalisers>

>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
>> trial. Simplify your report design, integration and deployment - and focus on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> Jamvm-general mailing list
>> Jamvm-general@...
>> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>>
>

MIPS

2009-11-16T03:37:47Z

Hello,

I am currently looking to port the 'inline' part of the latest 1.5.3 version to the MIPS platform so we can benefit from the performance increase inlining brings.

Does anyone have any information on how to do that? Or where to look for documentation on how this inlining part works?

It does not need to be a ready made solution (although I wouldn't mind :-) ), but just a tip or hint would already help me a lot.

Thanks

Re: Jamvm Performance

2009-11-12T09:59:05Z

Adding -verbosegc to the command line will show when a gc occurs, and
how long it takes.

Rob

2009/11/12 Arigead <captain.deadly@...>:

> David Griffiths wrote:
>> I believe JamVM uses a "stop-the-world" Garbage Collector, so called
>> because when you run out of heap it has to stop all threads while the
>> Garbage Collector (GC) runs. The GC pause can be quite long depending
>> on how many objects are in the heap. There are other JVMs out there
>> that minimise this pause through the use of more sophisticated GC
>> algorithms but at the cost of increased complexity and footprint.
>>
>> Cheers,
>>
>> Dave
>>
>> On Thu, Nov 12, 2009 at 3:01 PM, Arigead <captain.deadly@...> wrote:
>>> Hi David,
>>> Sorry about this, new to Java, well this indept part of java but
>>> what is the GC Cycle?
>>
>
> I should have asked in that previous email is there any way that I can
> enable some debug in the Jamvm build I have to see if this is the
> problem. I'm building from source so if there's debug I can easily
> switch it on if that'll shed some light onto things. Of course the
> overhead of debug might not help me.
>
> Thanks for your help
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

Re: Jamvm Performance

2009-11-12T09:45:21Z

David Griffiths wrote:

> I believe JamVM uses a "stop-the-world" Garbage Collector, so called
> because when you run out of heap it has to stop all threads while the
> Garbage Collector (GC) runs. The GC pause can be quite long depending
> on how many objects are in the heap. There are other JVMs out there
> that minimise this pause through the use of more sophisticated GC
> algorithms but at the cost of increased complexity and footprint.
>
> Cheers,
>
> Dave
>
> On Thu, Nov 12, 2009 at 3:01 PM, Arigead <captain.deadly@...> wrote:
>> Hi David,
>> Sorry about this, new to Java, well this indept part of java but
>> what is the GC Cycle?
>

I should have asked in that previous email is there any way that I can
enable some debug in the Jamvm build I have to see if this is the
problem. I'm building from source so if there's debug I can easily
switch it on if that'll shed some light onto things. Of course the
overhead of debug might not help me.

Thanks for your help

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Jamvm Performance

2009-11-12T09:42:57Z

David Griffiths wrote:

thanks David,
given my code is in an osgi bundle running in the Knophleurfish osgi
container there may well be a lot of garbage to collect. That might well
be my issue. :-(

John

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Jamvm Performance

2009-11-12T07:52:36Z

I believe JamVM uses a "stop-the-world" Garbage Collector, so called
because when you run out of heap it has to stop all threads while the
Garbage Collector (GC) runs. The GC pause can be quite long depending
on how many objects are in the heap. There are other JVMs out there
that minimise this pause through the use of more sophisticated GC
algorithms but at the cost of increased complexity and footprint.

Cheers,

Dave

On Thu, Nov 12, 2009 at 3:01 PM, Arigead <captain.deadly@...> wrote:
> Hi David,
> Sorry about this, new to Java, well this indept part of java but
> what is the GC Cycle?

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Jamvm Performance

2009-11-12T07:01:45Z

Hi David,
Sorry about this, new to Java, well this indept part of java but
what is the GC Cycle? I suppose I should head to the Internet and learn
something. I'm off line at the moment so that'll have to wait until
morning. This email will go in the morning and I might have found out on
wikipedia by then.

Anyhow I should add to my previous email I've a few problems but the one
I have been facing is not a problem with Jamvm, well not as yet. ;-)
I've got knophlerfish osgi container running with a bundle of Java code
that listens to DBus signals (Simply play and pause from RhythmBox on
desktop of eee). At the same time as running that bundle in the osgi
container I have the identical code running as a stand alone Java App.
Both running on Jamvm and Classpath. In this set up the Standalone app
is receiving all the events but for some reason the osgi bundle doesn't.
All work fine with Sun Java 6 so there is something about that
combination of the osgi container and Jamvm.

I know that'll have to debug this somehow but where to start? Jamvm or
the osgi container.

David Griffiths wrote:

> Maybe it's during a GC cycle?
>
> Cheers,
>
> Dave
>
> On Thu, Nov 5, 2009 at 9:09 AM, Arigead <captain.deadly@...> wrote:
>> Hello all,
>> I posted previous emails to this list on the subject of JamVM and
>> JXTA. Unfortunately that was a few weeks ago as I got distracted by the
>> work that pays the bills ;-)
>>
>> I should maybe explain what it is I'm trying to do. I'm eventually
>> hoping to run OSGi Bundles in a knophlerfish container on two embedded
>> devices. This is all for an EU Project on Personal Smart Spaces,
>> (whatever they are ;-)
>>
>> The networking on this project is taken care of my Jxta, maybe "taken
>> care of" is a bit strong there. Anyhow this project is intended for
>> mobile devices and for some reason the project has gone with Sun Java 6.
>>
>> Maybe that's enough of an intro. Part of the project is a AI system that
>> analyses users activity and tries to pre-empt what they will do by
>> learning. A simple example would be that if you're listening to music at
>> home and every time you receive a phone call on your mobile, that you
>> answer, you press pause on the music player. The system (Two devices
>> sharing info via Jxta) should learn this after a few calls and
>> automatically "press" pause for you.
>>
>> That's a basic example but it'll do. I'd like to get this SW running on
>> a gumstix and an OpenMoko phone. To start out though I took a eeePC and
>> installed Ubuntu 9.10 Netbook Remix into it. Then installed Classpath
>> 0.98 and JamVM 1.5.3, both from source code. I needed ecj and gcj
>> installed on the machine to get Classpath installed but then I removed
>> them and created a symbolic link of java to jamvm.
>>
>> That done I startup knophlerfish and run a very simple bundle which uses
>> Java-DBus to "listen" for play and pause DBus signals from RhythmBox.
>>
>> Now here's the question Is there any way that I can debug JamVM as I'm
>> missing occasional Play and Pause Signals. According to the system
>> monitor my CPU and memory aren't up to much so I don't think the system
>> is stretched at all. If anybody could offer me some advice I would be
>> most grateful. I think that if I'm missing signals on an 1.6GHz Intel
>> Atom, there's not much hope of getting this running effectively on the
>> OpenMoko's Arm.
>>
>> Sorry for the long email but I thought I should explain a bit. Thanks a
>> million again for any help anybody can offer.
>>
>> John

Re: jamvm in ARM linux

2009-11-10T05:53:29Z

Thanks for your quickly answers, but today my ARM board arrive long long away :(

When i have board again i try to view clock with your command (date) and say you if this worked.

However, i try to install jamvm since arm linux (without cross compiler). Is possible compile jamvm and classpath direct in arm linux without using cross compiler in other computer?

Thanks for all.

Jose.

--- El mar, 10/11/09, Robert Lougher <rob.lougher@...> escribió:

De: Robert Lougher <rob.lougher@...>
Asunto: Re: [Jamvm-general] jamvm in ARM linux
Para: "Jose VR" <c4_esp_21@...>
CC: "jamvm-general@..." <jamvm-general@...>
Fecha: martes, 10 de noviembre, 2009 11:58

As the error message says, check your system clock, e.g.

> date

This will probably be completely wrong. Any files created by the configure script will get this time. This causes the configure script to complain, because build dependencies on based on time, and if your clock is wrong, files may not be correctly (re-)built.

Set your clock!

Rob.

2009/11/10 Jose VR <c4_esp_21@...>

Hi:

I try to create a makefile in my ARMv5, to do this execute ./configure in jamvm folder but shell show me an error:

cheking for BSD-compatible instal... /usr/bin/install -c
cheking whether build environment is sane...configure: error: newly created file is older than distributed files!
Check your system clock

I have linux 2.6.26 with gcc-4.1.2 and gnueabi

I compile jamvm succesfully in a PC but in ARM not found. How i can solve this problem?

Thanks for all.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: jamvm in ARM linux

2009-11-10T02:58:26Z

As the error message says, check your system clock, e.g.

> date

This will probably be completely wrong. Any files created by the configure script will get this time. This causes the configure script to complain, because build dependencies on based on time, and if your clock is wrong, files may not be correctly (re-)built.

Set your clock!

Rob.

2009/11/10 Jose VR <c4_esp_21@...>

Hi:

I try to create a makefile in my ARMv5, to do this execute ./configure in jamvm folder but shell show me an error:

cheking for BSD-compatible instal... /usr/bin/install -c
cheking whether build environment is sane...configure: error: newly created file is older than distributed files!
Check your system clock

I have linux 2.6.26 with gcc-4.1.2 and gnueabi

I compile jamvm succesfully in a PC but in ARM not found. How i can solve this problem?

Thanks for all.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

jamvm in ARM linux

2009-11-10T01:45:15Z

Hi:

I try to create a makefile in my ARMv5, to do this execute ./configure in jamvm folder but shell show me an error:

cheking for BSD-compatible instal... /usr/bin/install -c
cheking whether build environment is sane...configure: error: newly created file is older than distributed files!
Check your system clock

I have linux 2.6.26 with gcc-4.1.2 and gnueabi

I compile jamvm succesfully in a PC but in ARM not found. How i can solve this problem?

Thanks for all.

Jamvm Performance

2009-11-05T01:09:43Z

Hello all,
I posted previous emails to this list on the subject of JamVM and
JXTA. Unfortunately that was a few weeks ago as I got distracted by the
work that pays the bills ;-)

I should maybe explain what it is I'm trying to do. I'm eventually
hoping to run OSGi Bundles in a knophlerfish container on two embedded
devices. This is all for an EU Project on Personal Smart Spaces,
(whatever they are ;-)

The networking on this project is taken care of my Jxta, maybe "taken
care of" is a bit strong there. Anyhow this project is intended for
mobile devices and for some reason the project has gone with Sun Java 6.

Maybe that's enough of an intro. Part of the project is a AI system that
analyses users activity and tries to pre-empt what they will do by
learning. A simple example would be that if you're listening to music at
home and every time you receive a phone call on your mobile, that you
answer, you press pause on the music player. The system (Two devices
sharing info via Jxta) should learn this after a few calls and
automatically "press" pause for you.

That's a basic example but it'll do. I'd like to get this SW running on
a gumstix and an OpenMoko phone. To start out though I took a eeePC and
installed Ubuntu 9.10 Netbook Remix into it. Then installed Classpath
0.98 and JamVM 1.5.3, both from source code. I needed ecj and gcj
installed on the machine to get Classpath installed but then I removed
them and created a symbolic link of java to jamvm.

That done I startup knophlerfish and run a very simple bundle which uses
Java-DBus to "listen" for play and pause DBus signals from RhythmBox.

Now here's the question Is there any way that I can debug JamVM as I'm
missing occasional Play and Pause Signals. According to the system
monitor my CPU and memory aren't up to much so I don't think the system
is stretched at all. If anybody could offer me some advice I would be
most grateful. I think that if I'm missing signals on an 1.6GHz Intel
Atom, there's not much hope of getting this running effectively on the
OpenMoko's Arm.

Sorry for the long email but I thought I should explain a bit. Thanks a
million again for any help anybody can offer.

John

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

JamVM 1.5.3 performance issue on ARM

2009-10-14T01:23:45Z

Hi,

I have been successfully testing JamVM + GNU Classpath 0.98 on ARM a couple of weeks now. However, it seems that JamVM suffers significant performance problems in my test board.

For example running Scimark2 gives the following results:

Composite Score: 0.754515649534562
FFT (1024): 0.43564193417098535
SOR (100x100): 1.3841795476602183
Monte Carlo : 0.2546323439119081
Sparse matmult (N=1000, nz=5000): 0.7180925550940853
LU (100x100): 0.9800318668356128

In contrast to proprietary VMs such as Sun CDC HotSpot and HotSpot client VM, the performance is significantly slower. HotSpot client VM had composite score of 20 with JIT and 2,46 when running interpreter mode.

I have used both direct-threaded interpreter with stack-caching and inline-threaded interpreter with stack-caching. Currently, I don't have a cross compiler with hard-float support so I'm only compiling JamVM with soft-float. Of course hard-float could improve the performance in numerical computing, but still, the performance is also quite bad when running for example a simple OSGi core implementation such as equinox. Startup takes approximately 1 minute.

The board has a ARM1136JF-S core.

Is this just an ARM issue with JamVM or am I having some other problems for example with JamVM features?

Very best regards,
Ronem

ThreadInfoonstructor not matching gnu classpath

2009-10-08T09:37:33Z

Hi!

Noticed something that seems to be missing in jamvm integration with gnu classpath.

jamvm 1.5.3 (and HEAD) is using a constructor for java.lang.management.ThreadInfo that was present in gnu classpath 0.93 but not in 0.95 and later. It has changed since then.

Line 1340 and 1374 in natives.c for jamvm 1.5.3.

best regards

Mattias

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: classpath - 0.97.2-r8

2009-10-02T06:12:51Z

Robert Lougher wrote:

> Hi Arigead,
>
> 2009/10/2 Arigead <captain.deadly@...>:
>> Hello all,
>> is there version requirement on classpath when using jamvm_1.5.3?
>> Tried to run something and got the following error. As yet have found
>> nothing searching on Internet for info on this error.
>>
>
> Yes, I've been wondering when you'd hit this problem :) Ugrading to
> JamVM 1.5.3 also requires updating GNU Classpath to 0.98 (the latest).
> I forgot to mention this, but the requirement is listed in both the
> README and INSTALL files.
>
> Hopefully this won't be very difficult to do. If you really get
> stuck, there is a way to add some compatibility classes so that JamVM
> 1.5.3 can work with Classpath 0.96/0.97, but it's a while since I've
> done this so I'll need to work out which ones again.
>
> Rob.
>

My bad sorry for disturbing you. Note to self read the readme now and
again. thanks for the heads up in spite of my omission in the reading
dept. ;-)

------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: classpath - 0.97.2-r8

2009-10-02T03:38:08Z

Hi Arigead,

2009/10/2 Arigead <captain.deadly@...>:
>
> Hello all,
> is there version requirement on classpath when using jamvm_1.5.3?
> Tried to run something and got the following error. As yet have found
> nothing searching on Internet for info on this error.
>

Yes, I've been wondering when you'd hit this problem :) Ugrading to
JamVM 1.5.3 also requires updating GNU Classpath to 0.98 (the latest).
I forgot to mention this, but the requirement is listed in both the
README and INSTALL files.

Hopefully this won't be very difficult to do. If you really get
stuck, there is a way to add some compatibility classes so that JamVM
1.5.3 can work with Classpath 0.96/0.97, but it's a while since I've
done this so I'll need to work out which ones again.

Rob.

> Exception in thread "main" java.lang.InternalError: Expected reflection
> field doesn't exist
> at java.lang.VMClass.getDeclaredConstructors(Native Method)
> at java.lang.Class.getDeclaredConstructors(Class.java:430)
> at java.lang.Class.newInstance(Class.java:1116)
> at java.util.logging.LogManager.createInstance(LogManager.java:854)
> at java.util.logging.LogManager.readConfiguration(LogManager.java:572)
> at java.util.logging.LogManager.readConfiguration(LogManager.java:527)
> at java.util.logging.LogManager.initLogManager(LogManager.java:203)
> at java.util.logging.LogManager.getLogManager(LogManager.java:168)
> at java.util.logging.Logger.getLogger(Logger.java:252)
> at java.util.logging.Logger.getLogger(Logger.java:211)
> at java.util.logging.Logger$1.run(Logger.java:94)
> at java.security.AccessController.doPrivileged(AccessController.java:96)
> at java.util.logging.Logger.<clinit>(Logger.java:90)
> at gnu.java.util.jar.JarUtils.<clinit>(JarUtils.java:65)
> at java.util.jar.Manifest.read(Manifest.java:162)
> at java.util.jar.Manifest.<init>(Manifest.java:89)
> at java.util.jar.JarFile.readManifest(JarFile.java:303)
> at java.util.jar.JarFile.<init>(JarFile.java:202)
> at java.util.jar.JarFile.<init>(JarFile.java:181)
> at jamvm.java.lang.JarLauncher.main(JarLauncher.java:33)
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry® Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9-12, 2009. Register now!
> http://p.sf.net/sfu/devconf
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

classpath - 0.97.2-r8

2009-10-02T03:17:28Z

Hello all,
is there version requirement on classpath when using jamvm_1.5.3?
Tried to run something and got the following error. As yet have found
nothing searching on Internet for info on this error.

Exception in thread "main" java.lang.InternalError: Expected reflection
field doesn't exist
at java.lang.VMClass.getDeclaredConstructors(Native Method)
at java.lang.Class.getDeclaredConstructors(Class.java:430)
at java.lang.Class.newInstance(Class.java:1116)
at java.util.logging.LogManager.createInstance(LogManager.java:854)
at java.util.logging.LogManager.readConfiguration(LogManager.java:572)
at java.util.logging.LogManager.readConfiguration(LogManager.java:527)
at java.util.logging.LogManager.initLogManager(LogManager.java:203)
at java.util.logging.LogManager.getLogManager(LogManager.java:168)
at java.util.logging.Logger.getLogger(Logger.java:252)
at java.util.logging.Logger.getLogger(Logger.java:211)
at java.util.logging.Logger$1.run(Logger.java:94)
at java.security.AccessController.doPrivileged(AccessController.java:96)
at java.util.logging.Logger.<clinit>(Logger.java:90)
at gnu.java.util.jar.JarUtils.<clinit>(JarUtils.java:65)
at java.util.jar.Manifest.read(Manifest.java:162)
at java.util.jar.Manifest.<init>(Manifest.java:89)
at java.util.jar.JarFile.readManifest(JarFile.java:303)
at java.util.jar.JarFile.<init>(JarFile.java:202)
at java.util.jar.JarFile.<init>(JarFile.java:181)
at jamvm.java.lang.JarLauncher.main(JarLauncher.java:33)

------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Jamvm and Jxta

2009-09-24T17:33:39Z

Hi John,

If you do:

jamvm -version

this will show the Java compatible version and the JamVM version,
which is what I'm interested in, e.g.:

java version "1.5.0"
JamVM version 1.5.3

...

Anyway, I found your thread here:

http://forums.java.net/jive/thread.jspa?threadID=66637&tstart=-1

One of the replies indicates that JXTA is doing a check involving:

Package.getPackage("java.lang").getSpecificationVersion()

which was returning NULL on their VM (from IBM).

On JamVM, this will also return NULL on versions before 1.5.3. But it
works correctly on the lastest version:

class foo {
static void main(String[] args) {
System.out.println(Package.getPackage("java.lang").getSpecificationVersion());
}
}

> jamvm -version

java version "1.5.0"
JamVM version 1.5.3

> jamvm foo
1.5

> jamvm -version

java version "1.5.0"
JamVM version 1.5.1

> jamvm foo

java.lang.NullPointerException
at foo.main(foo.java:3)

Hope this helps,
Rob.

2009/9/24 Arigead <captain.deadly@...>:

> Hi Robert thanks for getting back to me I'm running the bitbake recipe
> from Openembedded which builds version 1.5
>
> I'm fairly sure Jxta working with SUN Java 1.5
>
> John
>
>
> Robert Lougher wrote:
>> Hi,
>>
>> Which version of JamVM are you using?
>>
>> Rob.
>>
>> 2009/9/24 Arigead <captain.deadly@...>:
>>> Hello all,
>>> first post to this list, just signed up last night as I'm having a
>>> problem with running Jxta on Jamvm. I'm trying to run Jxta on a gumstix
>>> embedded system, amongst others but hit a problem when Jxta tries to
>>> call java.lang.IsCompatibleWith()
>>>
>>> I suppose I could try and get a hold of the jamvm source and investigate
>>> why I'm getting a null pointer exception. If anybody has any ideas
>>> please let me know as Java ain't my forte at all.
>>>
>>> Thanks in advance
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Come build with us! The BlackBerry® Developer Conference in SF, CA
>>> is the only developer event you need to attend this year. Jumpstart your
>>> developing skills, take BlackBerry mobile applications to market and stay
>>> ahead of the curve. Join us from November 9-12, 2009. Register now!
>>> http://p.sf.net/sfu/devconf
>>> _______________________________________________
>>> Jamvm-general mailing list
>>> Jamvm-general@...
>>> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>>>
>>
>
>

Re: Jamvm and Jxta

2009-09-24T02:59:52Z

Hi,

Which version of JamVM are you using?

Rob.

2009/9/24 Arigead <captain.deadly@...>:

> Hello all,
> first post to this list, just signed up last night as I'm having a
> problem with running Jxta on Jamvm. I'm trying to run Jxta on a gumstix
> embedded system, amongst others but hit a problem when Jxta tries to
> call java.lang.IsCompatibleWith()
>
> I suppose I could try and get a hold of the jamvm source and investigate
> why I'm getting a null pointer exception. If anybody has any ideas
> please let me know as Java ain't my forte at all.
>
> Thanks in advance
>
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry® Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9-12, 2009. Register now!
> http://p.sf.net/sfu/devconf
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

Jamvm and Jxta

2009-09-24T02:52:29Z

Hello all,
first post to this list, just signed up last night as I'm having a
problem with running Jxta on Jamvm. I'm trying to run Jxta on a gumstix
embedded system, amongst others but hit a problem when Jxta tries to
call java.lang.IsCompatibleWith()

I suppose I could try and get a hold of the jamvm source and investigate
why I'm getting a null pointer exception. If anybody has any ideas
please let me know as Java ain't my forte at all.

Thanks in advance

------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Performance issue

2009-09-21T06:07:12Z

Hi,

On Fri, Sep 18, 2009 at 7:23 PM, Robert Lougher <rob.lougher@...> wrote:
> Hi,
>
> Some questions:
>
> 1) What is a at91?
ARM920T around 150Mhz, 64MB of memory
http://www.atmel.com/dyn/products/datasheets.asp?family_id=605
AT91RM9200 Series

> 2) What version of sun-cdc were you using?
I don't know, I suspect a SUN embedded with hotspot(a commercial version).

> 4) What VMs were you using on the PC?
Sun JSE 6.0 vs JamVM.

> 5) What is the output of jamvm -version?
Execution Engine: direct-threaded interpreter with stack-caching

>
> Try increasing -Xms and -Xmx (e.g. -Xms16m). By default these values
> are very low on ARM (especially the initial heap setting -Xms). It
done, but not better.

> could be spending a lot of time garbage collecting (do -Xverbose:gc to
> see).
Ok next week for a test

Claude

>
> The code-copying JIT is not on by default on ARM, while it is enabled
> on desktop systems. But a difference of 12 is far, far too big,
> especially as sun-cdc is not as optimised as Hotspot.
>
> Rob.
>
> 2009/9/18 Claude Vittoria <claude.vittoria@...>:
>> Hi,
>>
>> I done some tests, with knopflerfish, jam1.5.3 and GNU Classpath09.8 on a at91.
>> The test is the web service of knopper which provids huge web pages,
>> many CSS and imgs.
>> I compare with sun-cdc, jamvm is 12time slower... I don't understand
>> the issue. I try to profile but I don't how to do. On a PC I get only
>> a difference of 3.
>>
>> How to explore the issue ? I don't know how to do to find if the issue
>> is in jam or in classpath.
>> What are the good option to compile jam on ARM ?
>>
>> Cheers,
>> Claude
>>
>> ------------------------------------------------------------------------------
>> Come build with us! The BlackBerry® Developer Conference in SF, CA
>> is the only developer event you need to attend this year. Jumpstart your
>> developing skills, take BlackBerry mobile applications to market and stay
>> ahead of the curve. Join us from November 9-12, 2009. Register now!
>> http://p.sf.net/sfu/devconf
>> _______________________________________________
>> Jamvm-general mailing list
>> Jamvm-general@...
>> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>>
>

Re: Performance issue

2009-09-18T10:23:02Z

Hi,

Some questions:

1) What is a at91?
2) What version of sun-cdc were you using?
4) What VMs were you using on the PC?
5) What is the output of jamvm -version?

Try increasing -Xms and -Xmx (e.g. -Xms16m). By default these values
are very low on ARM (especially the initial heap setting -Xms). It
could be spending a lot of time garbage collecting (do -Xverbose:gc to
see).

The code-copying JIT is not on by default on ARM, while it is enabled
on desktop systems. But a difference of 12 is far, far too big,
especially as sun-cdc is not as optimised as Hotspot.

Rob.

2009/9/18 Claude Vittoria <claude.vittoria@...>:

> Hi,
>
> I done some tests, with knopflerfish, jam1.5.3 and GNU Classpath09.8 on a at91.
> The test is the web service of knopper which provids huge web pages,
> many CSS and imgs.
> I compare with sun-cdc, jamvm is 12time slower... I don't understand
> the issue. I try to profile but I don't how to do. On a PC I get only
> a difference of 3.
>
> How to explore the issue ? I don't know how to do to find if the issue
> is in jam or in classpath.
> What are the good option to compile jam on ARM ?
>
> Cheers,
> Claude
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry® Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9-12, 2009. Register now!
> http://p.sf.net/sfu/devconf
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

Performance issue

2009-09-18T10:12:32Z

Hi,

I done some tests, with knopflerfish, jam1.5.3 and GNU Classpath09.8 on a at91.
The test is the web service of knopper which provids huge web pages,
many CSS and imgs.
I compare with sun-cdc, jamvm is 12time slower... I don't understand
the issue. I try to profile but I don't how to do. On a PC I get only
a difference of 3.

How to explore the issue ? I don't know how to do to find if the issue
is in jam or in classpath.
What are the good option to compile jam on ARM ?

Cheers,
Claude

------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

UnsatisfiedLinkError JamVM1.5.3 GNU-Classpath 0.98-1 on mips

2009-09-05T13:08:02Z

Hi,
I'm trying to use JamVM in an Asus wl-500w with openWrt 8.09.
I install GNU Classpath from http://ipkg.nslu2-linux.org/feeds/optware/openwrt-brcm24/cross/unstable/
which version is 0.98-1 and install the JamVM from repository.

I put all things in /mnt/opt/
I make a soft links from /mnt/opt/ to /opt and put /mnt/opt/bin to PATH.

When I launch jamvm -version
java version "1.5.0"
JamVM version 1.5.3
Copyright (C) 2003-2009 Robert Lougher <rob@lougher.org.uk>
/***
Same info
****/
Build information:

Execution Engine: direct-threaded interpreter with stack-caching
Compiled with: gcc 3.4.6 (OpenWrt-2.0)

Boot Library Path: /opt/lib/classpath
Boot Class Path: /opt/share/jamvm/classes.zip:/opt/share/classpath/glibj.zip

But when I try to run a helloWorld it throw me this:
root@OpenWrt:/mnt# jamvm HelloWorld
Exception occurred while printing exception (java/lang/NoClassDefFoundError)...
Original exception was java/lang/UnsatisfiedLinkError

I don't have idea what it could be. I have attached (HelloWorld.stack) the full verbose
when launching jamvm -verbose HelloWorld.

Someone has suffered from something similar?
Thanks
Ken

Tools.zip in bootclasspath by default

2009-09-02T02:36:41Z

Hi,

I don't find the configure option to get tools.zip in Boot Class Path
by default, a hint ?

Thanks,
Claude

===================================

java version "1.5.0"
JamVM version 1.5.3
Copyright (C) 2003-2009 Robert Lougher <rob@...>

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2,
or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.

Build information:

Execution Engine: inline-threaded interpreter with stack-caching
Compiled with: gcc 4.3.3

Boot Library Path: /home/next/rr/lib/classpath
Boot Class Path:
/home/next/rr/share/jamvm/classes.zip:/home/next/rr/share/classpath/glibj.zip

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

[patch] FreeBSD/sparc64 support

2009-08-30T02:03:32Z

Hello,

The attached patch adds support for FreeBSD/sparc64 without stack
caching or inlining to jamvm-1.5.3. The patch includes the files
generated by autoconf.

libffi is required.

gcMemMalloc/gcMemReallc/gcMemFree now stores the size of the allocated
area in an uintptr_t to avoid unaligned access traps.

Floats are now stored in the lower 32 bits of a stack slot (same place
as ints are already stored on 64 bit big endian machines) so that
get/setfield operations work. I think this must have been a problem on
ppc64 as well, but I have no machine to test it on. The symptom was an
OutOfMemoryError on startup.

Due to the above change callNative64.S for ppc64 must be updated to
add an offset of 4 when calling a function with float arguments.

Configure chooses -lthr over -lpthread on FreeBSD. pthread didn't work
at all on sparc64.
Changed the test for __thread so that it fails for FreeBSD/sparc64,
which it should.

A loop in inlining.c used a deleted piece of memory.

fmodf is used instead of fmod in the FREM instruction. This is easily
changed back if not correct.

A question: Is it hard to bring back the possibility to use
classpath:s older than 0.98?
Classpath 0.93 is the last one that can be compiled with jikes, all
the later versions need java. I solved this by using a patched
jamvm-1.5.1 + ecj.jar, but it is not very practical.

BR, pi

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

jamvm153_sparc64_diff.txt.gz (17K) Download Attachment

Re: JamVM 1.5.1 and Felix

2009-08-29T14:34:10Z

Hi,

2009/8/29 Claude Vittoria <claude.vittoria@...>:

>
> I redone my tests the compilers 4.3.2(debian) and 4.3.3(Ubuntu, debian
> sid) produce the issue
> 4.1.3 and 4.2.2 are ok.
> I done the tests with a classpath compiled with gcc3.4 and with the
> same version of gcc which compiled JamVM. Classpath doesn't seem to be
> affected by gcc version.
>
> Each time (jamvm -jar bin/felix.jar) if the error message
> *Couldn't find Main-Class attribute in bin/felix.jar Manifest*
> is displayed. The -cp path/to/class gives a deadlock VM.
>
> the configures:
> ../../classpath-0.98/configure --prefix=/home/cvittoria/trenext/usr
> --disable-gtk-peer --disable-gconf-peer --disable-plugin
> --disable-alsa
> ../../jamvm-1.5.1/configure --prefix=/home/cvittoria/trenext/usr
> --with-classpath-install-dir=/home/cvittoria/trenext/usr
>

Thanks for clarifying the gcc versions. Aliasing changes in gcc 4.3
broke JamVM 1.5.0 and 1.5.1. This was fixed in version 1.5.2. From
the NEWS file:

- Strict-aliasing fixes when compiled with GCC 4.3, interpreter
stack-caching is enabled, and interpreter inlining is on

Rob.

> Cheer,
> Claude
>

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: JamVM 1.5.1 and Felix

2009-08-29T14:22:45Z

Hi,

On Sat, Aug 29, 2009 at 8:49 PM, Robert Lougher<rob.lougher@...> wrote:

> Hi,
>
> 2009/8/29 Claude Vittoria <claude.vittoria@...>:
>> [snip]
>>
>> Ok after a lot of test and debug, I found the issue.
>> The problem is gcc 4.x, when I compile with gcc 3.4 JamVM does not get
>> a thread deadlock during loading classes.
>>
>
> Thanks for the report. A move from 1.x to 1.x+1 normally means major
> changes, and it sometimes takes a few releases to iron out the last
> bugs. I don't know if this is reason here but as it works with 1.5.3
> it is likely that the bug was introduced in 1.5.0, subsequently fixed.
>
>> arch x86
>> JamVM 1.5.1
>> Classpath 0.98
>> JDK ecj v_774_R33x, 3.3.1
>> Felix 1.8.0
>>
>
> Which version of gcc 4.x you are using, and what Linux distribution?
>
> Thanks,
> Rob.
>
>> I hope this might help someone,
>> Claude

I redone my tests the compilers 4.3.2(debian) and 4.3.3(Ubuntu, debian
sid) produce the issue
4.1.3 and 4.2.2 are ok.
I done the tests with a classpath compiled with gcc3.4 and with the
same version of gcc which compiled JamVM. Classpath doesn't seem to be
affected by gcc version.

Each time (jamvm -jar bin/felix.jar) if the error message
*Couldn't find Main-Class attribute in bin/felix.jar Manifest*
is displayed. The -cp path/to/class gives a deadlock VM.

the configures:
../../classpath-0.98/configure --prefix=/home/cvittoria/trenext/usr
--disable-gtk-peer --disable-gconf-peer --disable-plugin
--disable-alsa
../../jamvm-1.5.1/configure --prefix=/home/cvittoria/trenext/usr
--with-classpath-install-dir=/home/cvittoria/trenext/usr

Cheer,
Claude

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: JamVM 1.5.1 and Felix

2009-08-29T11:49:48Z

Hi,

2009/8/29 Claude Vittoria <claude.vittoria@...>:
> [snip]
>
> Ok after a lot of test and debug, I found the issue.
> The problem is gcc 4.x, when I compile with gcc 3.4 JamVM does not get
> a thread deadlock during loading classes.
>

Thanks for the report. A move from 1.x to 1.x+1 normally means major
changes, and it sometimes takes a few releases to iron out the last
bugs. I don't know if this is reason here but as it works with 1.5.3
it is likely that the bug was introduced in 1.5.0, subsequently fixed.

> arch x86
> JamVM 1.5.1
> Classpath 0.98
> JDK ecj v_774_R33x, 3.3.1
> Felix 1.8.0
>

Which version of gcc 4.x you are using, and what Linux distribution?

Thanks,
Rob.

> I hope this might help someone,
> Claude
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now. http://p.sf.net/sfu/bobj-july
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

Re: JamVM 1.5.1 and Felix

2009-08-29T05:20:13Z

[snip]

Ok after a lot of test and debug, I found the issue.
The problem is gcc 4.x, when I compile with gcc 3.4 JamVM does not get
a thread deadlock during loading classes.

arch x86
JamVM 1.5.1
Classpath 0.98
JDK ecj v_774_R33x, 3.3.1
Felix 1.8.0

I hope this might help someone,
Claude

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

JamVM 1.5.1 and Felix

2009-08-28T03:56:37Z

Hi,

I attempt to work with JamVM 1.5.1 and classpath 0.97 or 0.98 and
felix 1.8 or 1.0.4.

In all combinations with JamVM 1.5.1, I got this message :

$ export BOOTCLASSPATH=/home/next/trenext/jamvm-1.5.1/lib/classes.zip:/home/next/trenext/build/classpath/lib/glibj.zip:/home/next/trenext/build/classpath/tools/tools.zip
$ jamvm -verbose:class -jar bin/felix.jar
Couldn't find Main-Class attribute in bin/felix.jar Manifest.

If I try
$ jamvm -verbose:class -cp bin/felix.jar org.apache.felix.main.Main
I hang on [Loaded org/osgi/framework/FrameworkEvent]
The link of this class is not done and I never get the Felix shell.

I got a success, without manifest error message, with JamVM 1.5.3 and
classpath 0.98 and felix 1.8 or with JamVM 1.4.5 and classpath 0.97
and felix 1.0.3.

Why this combination JamVM 1.5.1 and classpath 0.97 or 0.98 and felix
1.8 or 1.0.4 does not work ?

Best Regards,
Claude

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Dalvik VM

2009-08-25T01:51:49Z

Hi Rob, have you thought about converting JamVM to use the Dalvik instruction set? It is supposed to be designed for low memory footprint and Dalvik is (I believe) interpreter only, so appears to target similar territory to JamVM.

Cheers,

Dave

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Re: Does SSL works at all in JamVm with classpath version 0.98 ???

2009-07-16T06:42:49Z

Hi,

It sounds as if you've found quite a few problems with the SSL
implementation in GNU Classpath. However, I know nothing about this
code. I just do VMs.

As you're not hitting problems with JamVM itself, I'd normally suggest
posting to the GNU Classpath mailing list instead (classpath@...).
But this has been very quiet of late. However, it's still your best
bet...

Rob.

2009/7/16 alk.shr <alk.shr@...>:

>
> i am not able to run my web application on HTTPS port (secure port through
> SSL) although it run fine on normal HTTP port.
>
> Environment used to run web application is as-
> GNU Classpath ver 0.98
> JamVm ver 1.53
> Jetty 1.6.8
> Linux Debian based
> IE and mozilla browser
> keystore- type GKR
> Crypto and SSL implementation- in GNU Classpath ver 0.98
>
> Basically i got multiple issues while trying to run application on HTTPS
> ports, some of them i was able to resolve after debugging GNU classpath
> source code. Problems faced by me are described below-
>
> Problem # 1
> ----------------------
> -Server socket listening on HTTPS port (8443 in our case) not responding to
> requests coming from browser.
>
> After analysis, i have found that SSL server socket has been listening on
> HTTPS port 8443 and accepting initial request coming from browser for
> connection and creating SSL client socket in response. but after this there
> is no response from SSL client socket created earlier. It seem that no input
> stream is open to the client socket to read data coming from browser.
>
>
> I think above issue is coming due to some bug in the SSLSocketImpl class
> under gnu.javax.net.ssl.provider package. In constructor of this class,a new
> Socket is created (i do not know why??) which is stored in underlyingSocket
> variable of SSLSocketImpl class. All requests of read and write is then
> delegated to member variable underlyingSocket. I think after copying new
> socket reference to underlyingSocket variable, this socket(underlyingSocket)
> is not connected to same native socket which is created in response of
> initial request from browser therefore SSL client socket is not responding
> to the browser request.
>
> I have fixed this issue by not setting underlyingSocket variable to new
> Socket and adding check for null at all places where underlyingSocket is
> refereed. I have diverted all calls on underlyingSocket to super class of
> SSLSocketImpl.
>
> Please confirm whether is this a bug in the SSLSocketImpl class or have i
> done something wrong?
>
> Problem # 2
> -----------------------
> SSL handshake starts working but IllegalArgumentException exception is
> coming from setLength API in Record class under gnu.javax.net.ssl.provider
> package.
>
>
> i think length check (between 0 and 16384 (2^14)) on SSL record is not
> correct. As per SSL RFC, length of final SSL record after encryption and
> compression may exceed by 2048 bytes.
>
> I have fixed this issue by changing maximum length to 17408.
>
> Please confirm is this the bug in the Record class?
>
> Problem # 3
> ------------------------
> In decrypt API of InputSecurityParameters class under
> gnu.javax.net.ssl.provider package, sometimes length calculated for padding
> in case of block cipher is more then size of SSL record/fragment resulting
> in IllegalArgumentException.
>
> I have seen this issue only with Internet explorer browser. At line # 173 in
> this class, IllegalArgumentException comes on calling positing API of
> ByteBuffer due to passing negative index.
>
>
> else if (record.version().compareTo(ProtocolVersion.TLS_1) >= 0)
> {
> // In TLSv1 and later, the padding must be `padlen' copies of the
> // value `padlen'.
> byte[] pad = new byte[padlen];
>
> //IllegalArgumentException comes at below line
> ((ByteBuffer) fragment.duplicate().position(record.length() - padlen
> - 1)).get(pad);
>
>
> for (int i = 0; i < pad.length; i++)
> if ((pad[i] & 0xFF) != padlen)
> badPadding = true;
> if (Debug.DEBUG)
> logger.logv(Component.SSL_RECORD_LAYER, "TLSv1.x padding\n{0}",
> new ByteArray(pad));
> }
>
>
> To resolve this issue, time being i have put safe check of positive index
> before the line where exception is coming.
>
>
>
>
> Now even after resolving all above mentioned issues, sometimes bad
> certificate or not valid signature error is coming on browser on opening
> pages using HTTPS.
>
> I have to provide HTTPS support and now i am really stuck. please guide me
> in resolving SSL related issues.
> --
> View this message in context: http://www.nabble.com/Does-SSL--works-at-all-in-JamVm-with-classpath-version-0.98-----tp24515753p24515753.html
> Sent from the JamVM mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> Enter the BlackBerry Developer Challenge
> This is your chance to win up to $100,000 in prizes! For a limited time,
> vendors submitting new applications to BlackBerry App World(TM) will have
> the opportunity to enter the BlackBerry Developer Challenge. See full prize
> details at: http://p.sf.net/sfu/Challenge
> _______________________________________________
> Jamvm-general mailing list
> Jamvm-general@...
> https://lists.sourceforge.net/lists/listinfo/jamvm-general
>

------------------------------------------------------------------------------
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in prizes! For a limited time,
vendors submitting new applications to BlackBerry App World(TM) will have
the opportunity to enter the BlackBerry Developer Challenge. See full prize
details at: http://p.sf.net/sfu/Challenge
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general

Does SSL works at all in JamVm with classpath version 0.98 ???

2009-07-16T05:53:54Z

i am not able to run my web application on HTTPS port (secure port through SSL) although it run fine on normal HTTP port.

Environment used to run web application is as-
GNU Classpath ver 0.98
JamVm ver 1.53
Jetty 1.6.8
Linux Debian based
IE and mozilla browser
keystore- type GKR
Crypto and SSL implementation- in GNU Classpath ver 0.98

Basically i got multiple issues while trying to run application on HTTPS ports, some of them i was able to resolve after debugging GNU classpath source code. Problems faced by me are described below-

Problem # 1
----------------------
-Server socket listening on HTTPS port (8443 in our case) not responding to requests coming from browser.

After analysis, i have found that SSL server socket has been listening on HTTPS port 8443 and accepting initial request coming from browser for connection and creating SSL client socket in response. but after this there is no response from SSL client socket created earlier. It seem that no input stream is open to the client socket to read data coming from browser.

I think above issue is coming due to some bug in the SSLSocketImpl class under gnu.javax.net.ssl.provider package. In constructor of this class,a new Socket is created (i do not know why??) which is stored in underlyingSocket variable of SSLSocketImpl class. All requests of read and write is then delegated to member variable underlyingSocket. I think after copying new socket reference to underlyingSocket variable, this socket(underlyingSocket) is not connected to same native socket which is created in response of initial request from browser therefore SSL client socket is not responding to the browser request.

I have fixed this issue by not setting underlyingSocket variable to new Socket and adding check for null at all places where underlyingSocket is refereed. I have diverted all calls on underlyingSocket to super class of SSLSocketImpl.

Please confirm whether is this a bug in the SSLSocketImpl class or have i done something wrong?

Problem # 2
-----------------------
SSL handshake starts working but IllegalArgumentException exception is coming from setLength API in Record class under gnu.javax.net.ssl.provider package.

i think length check (between 0 and 16384 (2^14)) on SSL record is not correct. As per SSL RFC, length of final SSL record after encryption and compression may exceed by 2048 bytes.

I have fixed this issue by changing maximum length to 17408.

Please confirm is this the bug in the Record class?

Problem # 3
------------------------
In decrypt API of InputSecurityParameters class under gnu.javax.net.ssl.provider package, sometimes length calculated for padding in case of block cipher is more then size of SSL record/fragment resulting in IllegalArgumentException.

I have seen this issue only with Internet explorer browser. At line # 173 in this class, IllegalArgumentException comes on calling positing API of ByteBuffer due to passing negative index.

else if (record.version().compareTo(ProtocolVersion.TLS_1) >= 0)
{
// In TLSv1 and later, the padding must be `padlen' copies of the
// value `padlen'.
byte[] pad = new byte[padlen];

//IllegalArgumentException comes at below line
((ByteBuffer) fragment.duplicate().position(record.length() - padlen - 1)).get(pad);

for (int i = 0; i < pad.length; i++)
if ((pad[i] & 0xFF) != padlen)
badPadding = true;
if (Debug.DEBUG)
logger.logv(Component.SSL_RECORD_LAYER, "TLSv1.x padding\n{0}",
new ByteArray(pad));
}

To resolve this issue, time being i have put safe check of positive index before the line where exception is coming.

Now even after resolving all above mentioned issues, sometimes bad certificate or not valid signature error is coming on browser on opening pages using HTTPS.

I have to provide HTTPS support and now i am really stuck. please guide me in resolving SSL related issues.

Re: ExceptionInInitializerError w/1.5.3 + 0.98 on amd64

2009-07-06T17:56:47Z

Hi Rob,

First off, thank you for going the extra mile and doing an OpenBSD
install to investigate this. It is much appreciated.

Ahh, yes your diff does correct it. The problem is _BIG_ENDIAN is
always defined in endian.h on the BSD's. The following works for
the BSD's and should work for others as well:

-#if (defined(BYTE_ORDER) && (BYTE_ORDER == BIG_ENDIAN)) ||
defined(_BIG_ENDIAN)
+#if (defined(BYTE_ORDER) && (BYTE_ORDER == BIG_ENDIAN)) ||
(!defined(BYTE_ORDER) && defined(_BIG_ENDIAN))

Thanks again for the help.

Regards,
-Kurt

Robert Lougher wrote:

> Hi Kurt,
>
> I've set up an OpenBSD amd64 install and I've tracked down the
> problem. To fix it properly, I need to look at extending my
> endianness checks (which are failing on OpenBSD). However, to get it
> working patch src/interp/engine/interp.h as follows:
>
> --- interp.h-old 2009-07-07 01:19:16.000000000 +0100
> +++ interp.h 2009-07-07 01:19:31.000000000 +0100
> @@ -37,11 +37,7 @@
> #include <sys/param.h>
> #endif
>
> -#if (defined(BYTE_ORDER) && (BYTE_ORDER == BIG_ENDIAN)) || defined(_BIG_ENDIAN)
> -#define FLOAT_CONST(val) (uintptr_t)val << (sizeof(uintptr_t) * 8 - 32)
> -#else
> #define FLOAT_CONST(val) val
> -#endif
>
> #define FLOAT_1_BITS FLOAT_CONST(0x3f800000)
> #define FLOAT_2_BITS FLOAT_CONST(0x40000000)
>
> Rob.
>
> 2009/6/30 Kurt Miller <kurt@...>:
>> Hi Rob,
>>
>> Thanks for the help.
>>
>> Robert Lougher wrote:
>>> 2009/6/23 Kurt Miller <kurt@...>:
>>>> On OpenBSD -current i386 works great, but amd64 can't run anything at
>>>> the moment due to a problem initializing:
>>>>
>>>> Exception occurred while printing exception
>>>> (java/lang/NoClassDefFoundError)...
>>>> Original exception was java/lang/ExceptionInInitializerError
>>>>
>>>> Even the simplest HelloWorld fails to run on amd64. It appears to
>>>> a recent amd64 bug since 1.5.1 + 0.98 works great on amd64.
>>>>
>>> Looks like something's gone wrong in the initialisation of Classpath.
>>> A class static initialiser (<clinit>) has thrown an exception, which
>>> has lead to the class being marked as "bad".
>>>
>>>> I'm stuck at this point. I need some help on how to find the root
>>>> cause.
>>>>
>>> Usual suspects is failure to load Classpath's native libraries (e.g.
>>> libjava.so, etc.).
>> At this point in the initialization no dlopen's have been called on
>> classpath's native libraries yet.
>>
>>> As a matter of interest try -Xnoinlining. This
>>> will rule out any problems with the inlining code due to
>>> race-conditions on OpenBSD's different threading code (I've a
>>> suspicion there's one lurking somewhere).
>> -Xnoinlining didn't fix the problem. I still get the same failure.
>>
>> The following debug info shows the class static initializer that
>> is throwing the exception is java/lang/Runtime.
>>
>> Breakpoint 12, initClass (class=0x21010c4f8)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1289
>> 1289 state = CLASS_BAD;
>> (gdb) bt
>> #0 initClass (class=0x21010c4f8)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1289
>> #1 0x000000000041f925 in resolveMethod (class=0x21010b288, cp_index=225)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:202
>> #2 0x000000000042c113 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1480
>> #3 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x21010ba48,
>> mb=0x20bfa8000, jargs=0x7f7ffffd4820)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #4 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x21010ba48,
>> mb=0x20bfa8000)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #5 0x000000000040c98f in initClass (class=0x21010ba48)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> #6 0x000000000041fe18 in resolveField (class=0x21010b920, cp_index=14)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:319
>> #7 0x000000000042b772 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1321
>> ---Type <return> to continue, or q <return> to quit---
>> #8 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x21010b920,
>> mb=0x206d16c00, jargs=0x7f7ffffd4c50)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #9 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x21010b920,
>> mb=0x206d16c00)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #10 0x000000000040c98f in initClass (class=0x21010b920)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> #11 0x000000000041fe18 in resolveField (class=0x21010ac58, cp_index=35)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:319
>> #12 0x000000000042b772 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1321
>> #13 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x21010b288,
>> mb=0x20a24f1a0, jargs=0x7f7ffffd5080)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #14 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x21010b288,
>> mb=0x20a24f1a0)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #15 0x000000000040c98f in initClass (class=0x21010b288)
>> ---Type <return> to continue, or q <return> to quit---
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> #16 0x000000000041f925 in resolveMethod (class=0x21010b120, cp_index=10)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:202
>> #17 0x000000000042c113 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1480
>> #18 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x21010b120,
>> mb=0x20baf9400, jargs=0x7f7ffffd54c0)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #19 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x21010b120,
>> mb=0x20baf9400)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #20 0x000000000040c98f in initClass (class=0x21010b120)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> #21 0x000000000041f925 in resolveMethod (class=0x21010ad60, cp_index=66)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:202
>> #22 0x000000000042c113 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1480
>> #23 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x210109c80,
>> ---Type <return> to continue, or q <return> to quit---
>> mb=0x200c7c000, jargs=0x7f7ffffd5900)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #24 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x210109c80,
>> mb=0x200c7c000)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #25 0x000000000040c98f in initClass (class=0x210109c80)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> #26 0x000000000041f925 in resolveMethod (class=0x2101098a0, cp_index=301)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:202
>> #27 0x000000000042c113 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1480
>> #28 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x210109ab0,
>> mb=0x200c7fe00, jargs=0x7f7ffffd5d40)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #29 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x210109ab0,
>> mb=0x200c7fe00)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #30 0x000000000040c98f in initClass (class=0x210109ab0)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1267
>> ---Type <return> to continue, or q <return> to quit---
>> #31 0x000000000041fe18 in resolveField (class=0x2101098a0, cp_index=26)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/resolve.c:319
>> #32 0x000000000042b772 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:1321
>> #33 0x0000000000410ccf in executeMethodVaList (ob=0x0, class=0x2101098a0,
>> mb=0x207535bb0, jargs=0x7f7ffffd6170)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:101
>> #34 0x00000000004107d6 in executeMethodArgs (ob=0x0, class=0x2101098a0,
>> mb=0x207535bb0)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/execute.c:73
>> #35 0x000000000040d6a4 in getSystemClassLoader ()
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1571
>> #36 0x00000000004042dd in main (argc=2, argv=0x7f7ffffd63d0)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/jam.c:363
>> (gdb) frame 0
>> #0 initClass (class=0x21010c4f8)
>> at /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/class.c:1289
>> 1289 state = CLASS_BAD;
>> (gdb) p cb->name
>> $78 = 0x20bafa3e0 "java/lang/Runtime"
>>
>> Further debugging yields the cause is due to a NPE at line
>> 2185 in src/interp/engine/interp.c. I changed NULL_POINTER_CHECK
>> to segfault since it is the first NPE to occur:
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> [Switching to process 10136, thread 0x20405a800]
>> 0x000000000042d190 in executeJava ()
>> at
>> /usr/obj/amd64/ports/jamvm-1.5.3/jamvm-1.5.3/src/interp/engine/interp.c:2185
>> 2185 DEF_OPC_210(OPC_INVOKEVIRTUAL_QUICK, {
>>
>>
>> 2185 DEF_OPC_210(OPC_INVOKEVIRTUAL_QUICK, {
>> 2186 Class *new_class;
>> 2187
>> 2188 arg1 = ostack - INV_QUICK_ARGS(pc);
>> 2189 NULL_POINTER_CHECK(*arg1);
>> (gdb) p *arg1
>> $7 = 0
>> (gdb) p ostack
>> $8 = (uintptr_t *) 0x205a99738
>> (gdb) p pc->operand.uu.u1
>> $9 = 3
>> (gdb) p *pc
>> $10 = {handler = 0x42d165, operand = {u = 8592228355, i = 2293763, ii = {
>> i1 = 3, i2 = 35}, uu = {u1 = 3, u2 = 35}, uui = {u1 = 3, u2 = 35 '#',
>> i = 0 '\0'}, pntr = 0x200230003}}
>>
>>
>> At a loss at why *arg1 is NULL. Any ideas on what to pursue next?
>>
>> Regards,
>> -Kurt
>>

------------------------------------------------------------------------------
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in prizes! For a limited time,
vendors submitting new applications to BlackBerry App World(TM) will have
the opportunity to enter the BlackBerry Developer Challenge. See full prize
details at: http://p.sf.net/sfu/blackberry
_______________________________________________
Jamvm-general mailing list
Jamvm-general@...
https://lists.sourceforge.net/lists/listinfo/jamvm-general