[chaco] segfault with multiarray

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

[chaco] segfault with multiarray

Fred
Hi all,

I'm on a Debian box x86_64 (wheezy) with NumPy 1.6.2 & SciPy 0.10.0.

When I run the CME attached on my Debian box, I get a segfault with
multiarray.

When I run it on a squeeze/stable Debian box x86_64 (NumPy 1.4.1 & SciPy
0.7.2), I don't get segfault.

Any suggestion?

TIA


Cheers,

--
Fred

_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev

foo.tmp (822 bytes) Download Attachment
cme.py (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [chaco] segfault with multiarray

Robert Kern
On Mon, Jul 2, 2012 at 2:45 PM, Fred <[hidden email]> wrote:

> Hi all,
>
> I'm on a Debian box x86_64 (wheezy) with NumPy 1.6.2 & SciPy 0.10.0.
>
> When I run the CME attached on my Debian box, I get a segfault with
> multiarray.
>
> When I run it on a squeeze/stable Debian box x86_64 (NumPy 1.4.1 & SciPy
> 0.7.2), I don't get segfault.
>
> Any suggestion?

Can you provide the gdb backtrace? Given the use of memmap and the
large difference in numpy versions, I'm leaning towards numpy bug, but
the backtrace will help settle it.

--
Robert Kern
Enthought
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev
Reply | Threaded
Open this post in threaded view
|

Re: [chaco] segfault with multiarray

Warren Weckesser


On Mon, Jul 2, 2012 at 9:22 AM, Robert Kern <[hidden email]> wrote:
On Mon, Jul 2, 2012 at 2:45 PM, Fred <[hidden email]> wrote:
> Hi all,
>
> I'm on a Debian box x86_64 (wheezy) with NumPy 1.6.2 & SciPy 0.10.0.
>
> When I run the CME attached on my Debian box, I get a segfault with
> multiarray.
>
> When I run it on a squeeze/stable Debian box x86_64 (NumPy 1.4.1 & SciPy
> 0.7.2), I don't get segfault.
>
> Any suggestion?

Can you provide the gdb backtrace? Given the use of memmap and the
large difference in numpy versions, I'm leaning towards numpy bug, but
the backtrace will help settle it.


It might be http://projects.scipy.org/numpy/ticket/2148

Fred, try changing the mode of the memmap to 'r+' (i.e. make it writable, even thought you won't write to it).

Warren
 


--
Robert Kern
Enthought
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev


_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev
Reply | Threaded
Open this post in threaded view
|

Re: [chaco] segfault with multiarray

Fred
Le 02/07/2012 16:50, Warren Weckesser a écrit :

> It might be http://projects.scipy.org/numpy/ticket/2148
>
> Fred, try changing the mode of the memmap to 'r+' (i.e. make it
> writable, even thought you won't write to it).
Robert, Warren,

This tricks works fine.

Thanks a lot!



Cheers,

--
Fred
_______________________________________________
Enthought-Dev mailing list
[hidden email]
https://mail.enthought.com/mailman/listinfo/enthought-dev