Issue1190

on jython: 
import array
a=array.array('B') # unsigned char
print a.sizeitem   
give 2 on jython and 1 on cpython 2.5.2
I encounter this bug in tryng to run part of google appengine on jython

In our reading, this is implementation dependent. The actual size is now
1 byte (since we can play tricks with the underlying storage of
signed/unsigned), however, if we were to use itemsize=1, this causes
other code to fail because outside of the storage size, it can only be
properly be manipulated as a 2-byte short.

It's a sad news because it means i cant use google protobuf on jython !!!
i couldn't t find a workaround for it!

2009/3/8 Jim Baker <report@bugs.jython.org>

>
> Jim Baker <zyasoft@users.sourceforge.net> added the comment:
>
> In our reading, this is implementation dependent. The actual size is now
> 1 byte (since we can play tricks with the underlying storage of
> signed/unsigned), however, if we were to use itemsize=1, this causes
> other code to fail because outside of the storage size, it can only be
> properly be manipulated as a 2-byte short.
>
> ----------
> nosy: +zyasoft
> resolution:  -> wont fix
> status: open -> closed
>
> _______________________________________
> Jython tracker <report@bugs.jython.org>
> <http://bugs.jython.org/issue1190>
> _______________________________________
>

Probably the best course of action is to use the Java version of
Google's protocol buffers. There might be some additional wrapping
involved. Regardless, it's likely to be more efficient.

thanks for the tip

2009/3/21 Jim Baker <report@bugs.jython.org>

>
> Jim Baker <zyasoft@users.sourceforge.net> added the comment:
>
> Probably the best course of action is to use the Java version of
> Google's protocol buffers. There might be some additional wrapping
> involved. Regardless, it's likely to be more efficient.
>
> _______________________________________
> Jython tracker <report@bugs.jython.org>
> <http://bugs.jython.org/issue1190>
> _______________________________________
>