Re: Re: Python-3-Migration of existing Data.fs

On Tue, Mar 20, 2018 at 11:04 AM, Arthur LEDARD <[email protected]>
wrote:

> There it is : https://github.com/vert-rouge/sample-zodb-migration-2to3
> comments are in French, sorry about that.
>
> I would like to add that if you make a subtle change in ZODB._compat.py,
> my problem is solved.
>
> around line 50
>     class Unpickler(zodbpickle.pickle.Unpickler):
>         def __init__(self, f):
> -            super(Unpickler, self).__init__(f)
> +            super(Unpickler, self).__init__(f, errors="bytes")
>

I'm glad you have a workaround. :)

I don't like this as a general solution.  In fact, I'm not even sure it's a
proper local solution, as it seems to cause some oids to be deserialized as
bytes and some as unicode, although I bet something in the stack converts
back to bytes when a unicode oid is encountered,



>
>
> It’s done for loads around line 70
>     def loads(s):
>         return zodbpickle.pickle.loads(s, encoding='ASCII', errors='bytes')
>

That makes me uncomfortable for the same reason.

The problem is that we don't save oids as binary in persistent ids.  This
is fairly easy to fix.  A related problem is that we don't use protocol 3
under Python 2, which largely defeats zodbpickle.binary, because it causes
binary data to be treated as latin-1-encoded unicode. :(

Jim

-- 
Jim Fulton
http://jimfulton.info

-- 
You received this message because you are subscribed to the Google Groups "zodb" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.