[Python-Dev] New policies for the Derby -- please read!

[Terry Reedy]

On 1/25/2014 10:37 AM, Larry Hastings wrote:
> On 01/25/2014 07:26 AM, Nick Coghlan wrote:
>> However, you've indicated that adding varargs support is going to take
>> you quite a bit of work, so postponing it is an option definitely
>> worth considering at this point in the release cycle.
>
> It's worth considering.  I'm estimating it's about 1.5 days' worth of
> work.  Mainly because, at the same time, I need to teach Clinic to have
> separate namespaces for "parser" and "impl" functions.  At the same time
> I was going to implement a frequently-requested feature, allowing the C
> variable storing an argument to have a different name than the actual
> Python parameter.
>
> And it could be one of those "hey that was easier than I thought"
> things.  1.5 days is kind of worst-case.  So maybe the best thing would
> be to give it a half-day and see if it turned out to be easy.
>
> Of course, If we bring the Derby to a close now-ish (debate going on in
> python-committers right now!), then I'll definitely put it off until 3.5.

I have been annoyed by the mismatch between Python signatures and 
C-implementation for at least a decade. Now that the idea that all 
functions (except possible for range) should have Python signatures 
seems to have been accepted, I am willing for the full implementation to 
wait until 3.5. I think the Derby experiment has pretty well exposed 
what still needs to be done, so I expect it should be possible to have 
it all in 3.5 as long as people do not burn out first.

-- 
Terry Jan Reedy