On Sun, Jan 29, 2012 at 02:53:49AM -0800, Dima Pasechnik wrote:
> Here is an experiment I just tried on ARM:
> create linear code from the example in LinearCode? (it's the same as in
> pickle_jar.tar.bz2)
>
> Pickle it:
> sage: sage.structure.sage_object.picklejar(C,dir='/tmp/jj')
> Quit Sage.
>
> Start Sage and unpickle it:
> sage: sage.structure.sage_object.unpickle_all('/tmp/jj/')
> Successfully unpickled 1 objects.
> Failed to unpickle 0 objects.
>
> Now try unpickling this pickle on x86:
> sage: sage.structure.sage_object.unpickle_all('/tmp/jj/')
> * unpickle failure:
> load('/tmp/jj//_class__sage_coding_linear_code_LinearCode__.sobj')
> Failed:
> _class__sage_coding_linear_code_LinearCode__.sobj
> Successfully unpickled 0 objects.
> Failed to unpickle 1 objects.
>
>
> Thus, there is a genuine pickle/unpickle bug on ARM....
> (FYI, I've uploaded the pickle produced on ARM)
Displaying the actual unpickle exception, I get:
('value too large to convert to char', <built-in function
unpickle_matrix_mod2_dense_v1>, (4, 7, [137, 80, 78, 71, 13, 10, 26, 10, 0, 0,
0, 13, 73, 72, 68, 82, 0, 0, 0, 7, 0, 0, 0, 4, 1, 3, 0, 0, 0, 120, 8, 183, 62,
0, 0, 0, 6, 80, 76, 84, 69, 0, 0, 0, 255, 255, 255, 165, 217, 159, 221, 0, 0,
0, 16, 73, 68, 65, 84, 8, 153, 99, 144, 99, 72, 99, 88, 197, 160, 3, 0, 5, 2,
1, 91, 3, 248, 139, 52, 0, 0, 0, 0, 73, 69, 78, 68, 174, 66, 96, 130], 91))
So this is the same issue with char being unsigned on your platform.
Specifically, it seems the corresponding __reduce__ function converts a char
array into a list of python ints, which has range 0-255 for you, instead of the
-128-127 on x86. And the unpickle function has the opposite issue.
I wonder how many of these are around, and if there is any general fix for
them.
-Willem Jan
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
