> CC: [email protected]
>On 02/27/13 13:00, azurIt wrote:
>>> CC: [email protected]
>>> On 02/27/13 12:46, azurIt wrote:
>>>>> CC: [email protected]
>>>>> On 02/27/13 12:30, azurIt wrote:
>>>>>> Hm, strange, works also here on 2.6.6 (tried both 32bit and 64bit):
>>>>> Strange indeed. Michael or Azur, if the below test case is not working
>>>>> for you, could you help me in isolating what was failing before and how
>>>>> *actually* my patch fixes it?
>>>> yes, i will help you but what exaclty do you need? i can still reproduce
>>>> the problem on most of the servers (didn't patch them yet) on python 2.6.6
>>>> (32bit and 64bit).
>>> well, my stepping through the validation code showed me that
>>> Decimal('inf') > 6000.0 == False. The patch is to have float('inf') for
>>> Float and Double types because float('inf') > 6000.0 == True and
>>> type(6000.0) is float anyway.
>>>
>>> When Decimal('inf') > 6000.0 == false, the validate_native() calls
>>> retuns False, which in turn causes the ValidationError to be raised.
>>>
>>> What happens in your case? Why does the validation function return false?
>>
>> Can you hint me where exactly you put your tracing code? Thanks.
>
>
>Right here:
>https://github.com/arskom/spyne/blob/spyne-2.9.3/spyne/protocol/xml/model.py#L76
Ok, i probably got it. It's related to -inf in Python 2.6.6:
>>> Decimal('-inf') > 1.0
True
>>> Decimal('-inf') > 1
False
_______________________________________________
Soap mailing list
[email protected]
http://mail.python.org/mailman/listinfo/soap
