Open Stack

On 11/28/2014 12:06 AM, Ihar Hrachyshka wrote:
> On 27/11/14 12:09, Thomas Goirand wrote:
>> On 11/27/2014 12:31 AM, Donald Stufft wrote:
>>>
>>>> On Nov 26, 2014, at 10:34 AM, Thomas Goirand <zigo at debian.org>
>>>> wrote:
>>>>
>>>> Hi,
>>>>
>>>> I tried to package suds-jurko. I was first happy to see that
>>>> there was some progress to make things work with Python 3.
>>>> Unfortunately, the reality is that suds-jurko has many issues
>>>> with Python 3. For example, it has many:
>>>>
>>>> except Exception, e:
>>>>
>>>> as well as many:
>>>>
>>>> raise Exception, 'Duplicate key %s found' % k
>>>>
>>>> This is clearly not Python3 code. I tried quickly to fix some
>>>> of these issues, but as I fixed a few, others appear.
>>>>
>>>> So I wonder, what is the point of using suds-jurko, which is
>>>> half-baked, and which will conflict with the suds package?
>>>>
>>> It looks like it uses 2to3 to become Python 3 compatible.
> 
>> Outch! That's horrible.
> 
>> I think it'd be best if someone spent some time on writing real
>> code rather than using such a hack as 2to3. Thoughts anyone?
> 
> That sounds very subjective. If upstream is able to support multiple
> python versions from the same codebase, then I see no reason for them
> to split the code into multiple branches and introduce additional
> burden syncing fixes between those.
> 
> /Ihar

Objectively, using 2to3 sux, and it's much better to fix the code,
rather than using such a band-aid. It is possible to support multiple
version of Python with a single code base. So many projects are able to
do it, I don't see why suds would be any different.

Cheers,

Thomas