[Stackless] endian issues pickling tasklets

John Tantalo john.tantalo at gmail.com
Thu Nov 2 14:47:37 CET 2006


Do you know which part of the pickled tasklet would be machine
dependent? Would the solution be as simple as converting the string to
a byte stream, identifying the dependent sections, and flipping the
words?

- John

On 11/2/06, Richard Tew <richard.m.tew at gmail.com> wrote:
> On 11/1/06, John Tantalo <john.tantalo at gmail.com> wrote:
> > I'm trying to resume pickled tasklets across different architectures
> > (Intel, PPC, SPARC) and I'm running into a problem related to the
> > endian-ness of the platforms. PPC to and from SPARC works fine,
> > because they are all big endian, but I can't resume any pickled
> > tasklets on Intel because it is little endian. Is this a known bug?
> > Are there any workarounds? I'm using Python 2.4.3 Stackless 3.1b3
> > 060516 on each platform.
>
> Not so much a known bug, but rather expected behaviour I would
> think.  As such, I do not know of any workarounds.  In fact this is
> the first time I have heard someone mentioning experiencing it.
>
> My experience with how this pickling is implemented is limited only
> to debugging pickling problems when porting to later versions of
> Python, so I do not know enough to have any idea how to fix this
> or to work around it.
>
> But if you have any ideas or want to come up with a patch, it
> would be appreciated.
>
> Richard.
>

_______________________________________________
Stackless mailing list
Stackless at stackless.com
http://www.stackless.com/mailman/listinfo/stackless



More information about the Stackless mailing list