[pHash-support] Crash when adding to Java MVP tree

[Sebastian Gallese]

Good question.  I'm using the following terminal command to generate my crash:

java -Djava.library.path=/usr/local/lib pHash -mvp /home/kdge/mickeymouse
/home/kdge/edison/output0040.jpg

You can download the output0040.jpg image here ( http://senduit.com/416c91 )

The -mvp option, if I'm not mistaken, tells pHash to build an mvp tree
from the 1st argument (a filepath to a directory), and then query that
mvp tree with the 2nd argument (the file path to an image).

These are the java files I'm using ( the standard ones generated in
pHash-0.8.1/bindings/java ):
http://senduit.com/3e3fd0

The program DOES generate mvp files, but not sure if they are valid.
Here are the mvp files the above crash generates:
http://senduit.com/9d30fb

I modified the basic pHash.java file with some basic
System.out.println statements ( you can download the source
http://gist.github.com/316917 ).  Here are my terminal results:

kdge at kdge-laptop:~/pHash-0.8.1/bindings/java$ java
-Djava.library.path=/usr/local/lib pHash -mvp /home/kdge/mickeymouse
/home/kdge/edison/output0040.jpg
Instantiated new MVP tree!
Created array of MHImageHash!
#
# A fatal error has been detected by the Java Runtime Environment:

With this in mind, the code would have crashed at this line in the java:
boolean result = mvp.create(hashes);

Hope this helps.  Let me know if you need any other documentation or
information regarding my system/set-up.

Thanks,
Seb

On Sat, Feb 27, 2010 at 2:32 PM, Evan Klinger <eklinger at phash.org> wrote:
> Sebastian,
>
> Did this 7MB dataset cause a crash in the Java MVP functions? I was
> not able to reproduce it using the pHash.java example included in
> pHash.
>
> Evan
>
> On Sat, Feb 27, 2010 at 10:26 AM, Sebastian Gallese
> <sebastiangallese at gmail.com> wrote:
>> Here's my lasts crashes dataset at 7MB:
>> http://senduit.com/6986b8 (it should remain online for 1 week)
>>
>> If this works, I can send you about 4GBs of flickr images that caused
>> my first crash.
>>
>> Thanks,
>> Seb
>>
>> On Fri, Feb 26, 2010 at 4:43 PM, Evan Klinger <eklinger at phash.org> wrote:
>>> Sebastian,
>>>
>>> That is the correct way to build pHash with debug support.
>>> Unfortunately, the crash log doesn't contain much information other
>>> than the crash occurred in the ph_save_mvptree function in the native
>>> C++ code. How large is the image dataset? Is there any way we can
>>> obtain a copy to test with?
>>>
>>> Thanks
>>> Evan
>>>
>>> On Fri, Feb 26, 2010 at 11:35 AM, Sebastian Gallese
>>> <sebastiangallese at gmail.com> wrote:
>>>> I built pHash with the following.  Does this enable debug support?
>>>> sudo ./configure --enable-java --enable-debug
>>>>
>>>> I got this error when building an MVP file of about 500 images.  I've
>>>> successfully built an MVP tree with 40 images.
>>>>
>>>> kdge at kdge-laptop:~/pHash-0.8.1/bindings/java$ java
>>>> -Djava.library.path=/usr/local/lib pHash -mvp /home/kdge/mickeymouse
>>>> /home/kdge/edison/output0040.jpg
>>>> #
>>>> # A fatal error has been detected by the Java Runtime Environment:
>>>> #
>>>> #  SIGSEGV (0xb) at pc=0x9002e62d, pid=8966, tid=3065756528
>>>> #
>>>> # JRE version: 6.0_15-b03
>>>> # Java VM: Java HotSpot(TM) Server VM (14.1-b02 mixed mode linux-x86 )
>>>> # Problematic frame:
>>>> # C  [libpHash.so.0+0x862d]
>>>> #
>>>> # An error report file with more information is saved as:
>>>> # /home/kdge/pHash-0.8.1/bindings/java/hs_err_pid8966.log
>>>> #
>>>> # If you would like to submit a bug report, please visit:
>>>> #   http://java.sun.com/webapps/bugreport/crash.jsp
>>>> # The crash happened outside the Java Virtual Machine in native code.
>>>> # See problematic frame for where to report the bug.
>>>> #
>>>> Aborted
>>>>
>>>> Output of the log file:
>>>> http://gist.github.com/316064
>>>>
>>>> Thanks,
>>>> Seb
>>>>
>>>>
>>>> On Tue, Feb 16, 2010 at 11:47 AM, Evan Klinger <eklinger at phash.org> wrote:
>>>>> Sebastian,
>>>>> Did you compile pHash with debug support? If not, can you please try
>>>>> that and see if the crash log contains any more useful information?
>>>>> Thanks
>>>>>
>>>>> On Mon, Feb 15, 2010 at 11:26 PM, Sebastian Gallese
>>>>> <sebastiangallese at gmail.com> wrote:
>>>>>> Trying to build a MVP tree of 13215 jpgs, all of which run in the 100
>>>>>> to 300 kilobyte range.  While doing so, I ran into the following
>>>>>> error:
>>>>>>
>>>>>> # A fatal error has been detected by the Java Runtime Environment:
>>>>>> #
>>>>>> #  SIGSEGV (0xb) at pc=0x8fffa62d, pid=22963, tid=3065453424
>>>>>> #
>>>>>> # JRE version: 6.0_15-b03
>>>>>> # Java VM: Java HotSpot(TM) Server VM (14.1-b02 mixed mode linux-x86 )
>>>>>> # Problematic frame:
>>>>>> # C  [libpHash.so.0+0x862d]
>>>>>> #
>>>>>> # An error report file with more information is saved as:
>>>>>> # /home/kdge/workspace/sgallese/hs_err_pid22963.log
>>>>>> #
>>>>>> # If you would like to submit a bug report, please visit:
>>>>>> #   http://java.sun.com/webapps/bugreport/crash.jsp
>>>>>> # The crash happened outside the Java Virtual Machine in native code.
>>>>>> # See problematic frame for where to report the bug.
>>>>>> #
>>>>>>
>>>>>> The full log is available here: http://gist.github.com/305373
>>>>>>
>>>>>> Do you think this is a memory issue?  Should I try to augment the
>>>>>> amount of memory to the Java Virtual machine?
>>>>>> I had the process running for about 7 hours before it crashed.  I
>>>>>> think I might try to add the jpgs to the MVP tree one at a time and
>>>>>> see how that works.
>>>>>>
>>>>>> Thanks,
>>>>>> Seb
>>>>>> _______________________________________________
>>>>>> pHash-support mailing list
>>>>>> pHash-support at lists.phash.org
>>>>>> http://lists.phash.org/listinfo.cgi/phash-support-phash.org
>>>>>>
>>>>>
>>>> _______________________________________________
>>>> pHash-support mailing list
>>>> pHash-support at lists.phash.org
>>>> http://lists.phash.org/listinfo.cgi/phash-support-phash.org
>>>>
>>>
>> _______________________________________________
>> pHash-support mailing list
>> pHash-support at lists.phash.org
>> http://lists.phash.org/listinfo.cgi/phash-support-phash.org
>>
>