On Jul 19, 2010, at 11:52 AM, Jed Smith wrote:

> On Jul 19, 2010, at 10:35 AM, [email protected] wrote:
> 
>> 
>> On Jul 19, 2010, at 10:18 AM, Jed Smith wrote:
>> 
>>> On Jul 19, 2010, at 10:13 AM, [email protected] wrote:
>>> 
>>>> 
>>>> On Jul 19, 2010, at 10:04 AM, Jed Smith wrote:
>>>> 
>>>>> On Jul 18, 2010, at 12:25 AM, Tom Davis wrote:
>>>>> 
>>>>>> Judging by the exception, I guess they stopped implicitly casting 
>>>>>> variables used in format strings? That's not something I remember 
>>>>>> reading about in What's New, but it makes sense in that context. What 
>>>>>> header we're sending that isn't a string is another question.
>>>>> 
>>>>> That's eye-opening.
>>>>> 
>>>>> I always thought %s called str() on the arguments in question if they 
>>>>> weren't
>>>>> strings already.  If they did indeed change that, half of my code (not 
>>>>> only
>>>>> in libcloud) is going to break.  I'd hope that's a regression and not
>>>>> intentional...
>>>> 
>>>> Please see my previous message about the redundant header that's being 
>>>> explicitly set as an integer.
>>> 
>>> Oh, it's in the .join(values).  That makes more sense, and has never worked.
>>> 
>>> My apologies, read the exception wrong.
>> 
>> No problem, now, the question is who is going to remove the setting of the 
>> redundant header and get it committed?
> 
> I cast it to string as opposed to removing it.  There may be a reason it is 
> set
> that both of us are unaware of.  LIBCLOUD-43 created for tracking.
> 
> Fixed in r965538 = 04f1f94f1fe26d2ac36f29d7b7a57ea786fed070

As per a previous e-mail, it's set by httplib by default but I guess without 
testing, this is safer.

Thanks,

S

Reply via email to