regression: [MinGW] Slightly More Robust Stacktraces on Windows

regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Marco Trudel]

I refer to this patch:
http://gcc.gnu.org/ml/java-patches/2006-q3/msg00051.html

This breaks JNI in certain cases. The application just silently dies 
when trying to run a native method.
Unfortunately it is a foreign library, I do not have it's source code to 
track the problem down.
So Ranjit, do you have an idea what it could be?


Marco

Re: regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Ranjit Mathew]

On 10/6/06, Marco Trudel <mtrudel@gmx.ch> wrote:
> I refer to this patch:
> http://gcc.gnu.org/ml/java-patches/2006-q3/msg00051.html
>
> This breaks JNI in certain cases. The application just silently dies
> when trying to run a native method.
> Unfortunately it is a foreign library, I do not have it's source code to
> track the problem down.
> So Ranjit, do you have an idea what it could be?

Sorry, off the top of my head I don't. You will have to
debug this. If asked for a wild guess, I would say
that libffi doesn't have frame-pointers enabled.

Sorry I couldn't be of more help,
Ranjit.


-- 
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://rmathew.com/

Re: regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Marco Trudel]

Ranjit Mathew wrote:
> On 10/6/06, Marco Trudel <mtrudel@gmx.ch> wrote:
>> I refer to this patch:
>> http://gcc.gnu.org/ml/java-patches/2006-q3/msg00051.html
>>
>> This breaks JNI in certain cases. The application just silently dies
>> when trying to run a native method.
>> Unfortunately it is a foreign library, I do not have it's source code to
>> track the problem down.
>> So Ranjit, do you have an idea what it could be?
> 
> Sorry, off the top of my head I don't. You will have to
> debug this. If asked for a wild guess, I would say
> that libffi doesn't have frame-pointers enabled.
> 
> Sorry I couldn't be of more help,

Can we then revert that patch? I'm running out of time for my project 
and I needed already 2 days to track the problem down to that patch.
I tried reverting only that patch but it fails. Reverting backtrace.h to 
the previous revision (115069) doesn't work as well. My application 
doesn't even start (with the patch it dies when trying to load the lib 
in question)

Marco

Re: regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Marco Trudel]

I now got it working by reverting the latest three patches that were 
applied to backtrace.h.
The question now is what to do. I don't think we can leave it broken, do 
we? I currently don't have the time to debug and fix that myself, but I 
could provide you the failing example (it's one line of code).
Or is it better to remove the patch?


Marco


Marco Trudel wrote:
> Ranjit Mathew wrote:
>> On 10/6/06, Marco Trudel <mtrudel@gmx.ch> wrote:
>>> I refer to this patch:
>>> http://gcc.gnu.org/ml/java-patches/2006-q3/msg00051.html
>>>
>>> This breaks JNI in certain cases. The application just silently dies
>>> when trying to run a native method.
>>> Unfortunately it is a foreign library, I do not have it's source code to
>>> track the problem down.
>>> So Ranjit, do you have an idea what it could be?
>>
>> Sorry, off the top of my head I don't. You will have to
>> debug this. If asked for a wild guess, I would say
>> that libffi doesn't have frame-pointers enabled.
>>
>> Sorry I couldn't be of more help,
> 
> Can we then revert that patch? I'm running out of time for my project 
> and I needed already 2 days to track the problem down to that patch.
> I tried reverting only that patch but it fails. Reverting backtrace.h to 
> the previous revision (115069) doesn't work as well. My application 
> doesn't even start (with the patch it dies when trying to load the lib 
> in question)
> 
> Marco
> 
> 

Re: regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Ranjit Mathew]

On 10/6/06, Marco Trudel <mtrudel@gmx.ch> wrote:
> I now got it working by reverting the latest three patches that were
> applied to backtrace.h.
> The question now is what to do. I don't think we can leave it broken, do
> we? I currently don't have the time to debug and fix that myself, but I
> could provide you the failing example (it's one line of code).
> Or is it better to remove the patch?

I'm travelling right now (and will continue to for a month), so even
if you do send me the problematic code I can't do much with it. :-(
Sorry about that.

Thanks,
Ranjit.

> Marco
>
>
> Marco Trudel wrote:
> > Ranjit Mathew wrote:
> >> On 10/6/06, Marco Trudel <mtrudel@gmx.ch> wrote:
> >>> I refer to this patch:
> >>> http://gcc.gnu.org/ml/java-patches/2006-q3/msg00051.html
> >>>
> >>> This breaks JNI in certain cases. The application just silently dies
> >>> when trying to run a native method.
> >>> Unfortunately it is a foreign library, I do not have it's source code to
> >>> track the problem down.
> >>> So Ranjit, do you have an idea what it could be?
> >>
> >> Sorry, off the top of my head I don't. You will have to
> >> debug this. If asked for a wild guess, I would say
> >> that libffi doesn't have frame-pointers enabled.
> >>
> >> Sorry I couldn't be of more help,
> >
> > Can we then revert that patch? I'm running out of time for my project
> > and I needed already 2 days to track the problem down to that patch.
> > I tried reverting only that patch but it fails. Reverting backtrace.h to
> > the previous revision (115069) doesn't work as well. My application
> > doesn't even start (with the patch it dies when trying to load the lib
> > in question)
> >
> > Marco
> >
> >
>


-- 
Ranjit Mathew      Email: rmathew AT gmail DOT com

Bangalore, INDIA.    Web: http://rmathew.com/

Re: regression: [MinGW] Slightly More Robust Stacktraces on Windows

[Mohan Embar]

Hi Marco,

>I now got it working by reverting the latest three patches that were 
>applied to backtrace.h.

>The question now is what to do. I don't think we can leave it broken, do 
>we? I currently don't have the time to debug and fix that myself, but I 
>could provide you the failing example (it's one line of code).
>Or is it better to remove the patch?

If you can, create a Bugzilla entry for this. You never know.

-- Mohan
http://www.thisiscool.com/
http://www.animalsong.org/