see http://code.google.com/p/getpaid/source/detail?spec=svn2096&r=1315

1) I don't think the patch does what it intends to do
2) the way it is written kills unicode support for all Five formbase
based forms

in the code it says:

------------------

REQUEST INPUT PROCESSING

five destructively processes list and tuple inputs, when converting to
unicode
so we patch it to only happen once in the request.

-------------------

then the patch

-------------------------

from Products.Five.formlib.formbase import FiveFormlibMixin
from Products.Five.browser import decode

def update( self ):
    if getattr( self.request, '__inputs_processed', False ):
        decode.processInputs( self.request )
        decode.setPageEncoding( self.request )
        request.__inputs_processed = True
    super( FiveFormlibMixin, self).update()

FiveFormlibMixin.update = update

------------------------

my interpretation is that line 40 (http://code.google.com/p/getpaid/
source/browse/trunk/products/PloneGetPaid/_patch.py?
spec=svn2096&r=1315#40) should be:

if not getattr( self.request, '__inputs_processed', False ):

cheers
Matt

On Wed, Oct 22, 2008 at 10:18 PM, Matt Halstead <matt.halst...@gmail.com>wrote:

agreed, it certainly looks odd, and your change sounds good. the intent was
clearly that it should be executing once and only once, and that's the whole
purpose for it. the issue was on any z3 form processing a list it would,
attempt to decode a list of values twice, which occurred with multiple
formlib viewlets in the admin.  otoh, if people haven't been reporting any
issues with unicode list processing, or a decode error in particular, than
perhaps the whole patch can just be removed.

cheers,

kapil