Nor do I, and lol, that's exactly what I was considering doing.   
Thanks for getting to it first!  ;D

On the "support for Twisted 1.*", front, I've seen one "drop support  
for it" and no "please keep support for it".  I may give it a day or  
so but I'm definitely for dropping it.

Daniel

On Nov 14, 2005, at 7:00 PM, James Bunton wrote:

> This is getting interesting :P
>
> I don't want to drop support for such a new Twisted release.  
> Perhaps this little bit of fun I was experimenting with will be  
> useful for you too.
>
> Have a look at this file
> http://websvn.delx.cjb.net/websvn/filedetails.php? 
> repname=pymsnt&path=%2Ftrunk%2Fsrc%2Ftlib%2Fxmlw.py
>
> Basically it provides xmlw.Element, SuxElementStream, parseText and  
> parseFile
> It hopefully will work with Twisted 1.3, Twisted 2.0 with any words.
> It first checks for Twisted 2.x or greater, if not, it falls back  
> on our internal domish because of all those bugs :P
> It then tries to import the new twisted.words.xish, if that fails,  
> it imports from twisted.xish
>
> In all the other files we then import from tlib.xmlw.
>
> Sound ok?
>
> ---
>
> James
>
>
>
>
> On 15/11/2005, at 8:09 AM, Daniel Henninger wrote:
>
>> Ok, consider this fix a "hack" fix.  A real fix is to import the  
>> "correct" Element.  =/  Thing is, between 0.1.0 and 0.3.0 of  
>> words, a lot of the xish stuff seems to have made it's way in,  
>> instead of staying in Xish.  So now we have two versions of domish  
>> sitting around, one that's compatible with 0.1.0. and one that's  
>> compatible with 0.3.0.  Now... the below fix works fine, all  
>> things considered, because really, all that happens in the  
>> xmlstream code is that same thing ... if it's an Element, it's  
>> turned into a string with toXml().  So we're not hurting anything  
>> by doing it earlier, and that would provide backwards  
>> compatibility for now.  I believe I'm going to leave the hack in  
>> for now,  but it's looking like we're either going to need to  
>> embrace the new layout and say byebye to 0.1.0 support, or deal  
>> with it in a variety of other ways.
>>
>> Kind of a moving target sometimes, isn't it?  ;)
>>
>> Basically, the low down is that the xmlstream code checks that  
>> what's passed to send "isInstance" of domish.Element.  Well, it's  
>> not if we're not talking about the same exact domish.py.  So it  
>> was passing our Element on through, which is indeed not a "sized"  
>> element.
>>
>> Daniel
>>
>> On Nov 14, 2005, at 11:00 AM, Daniel Henninger wrote:
>>
>>> This is sort of for James's benefit, but I thought I would throw  
>>> it out to the list.  The change that occured that caused PyAIM to  
>>> fail with 0.3.0 ended up being that the send function no longer  
>>> accepts an Element.  It's expecting that you will hand it a  
>>> string.  I've been toying with overriding send anyway because I  
>>> need to do so for Jabberd2's component protocol (well, I think  I  
>>> haven't worked that functionality out yet).  So PyTransport need a:
>>> def send(self, obj):
>>>    if type(obj) == Element:
>>>       obj = obj.toXml()
>>>    component.Service.send(self,obj)
>>>
>>> That does the trick.  I'll be submitting the changes  
>>> momentarily.  I'm going to pose this change on the twisted jabber  
>>> list and see what is said.
>>>
>>> Daniel
>>>
>>> -- 
>>> "The most addictive drug in the world is music."
>>>      - The Lost Boyz
>>>
>>>
>>>
>>> _______________________________________________
>>> py-transports mailing list
>>> [email protected]
>>> http://www.modevia.com/cgi-bin/mailman/listinfo/py-transports
>>>
>>>
>>
>> _______________________________________________
>> py-transports mailing list
>> [email protected]
>> http://www.modevia.com/cgi-bin/mailman/listinfo/py-transports
>>
>
> _______________________________________________
> py-transports mailing list
> [email protected]
> http://www.modevia.com/cgi-bin/mailman/listinfo/py-transports
>
>

Reply via email to