[Pyrex] [Cython-dev] kw-only arguments?

[Robert Bradshaw]

On Sep 25, 2007, at 12:08 AM, Stefan Behnel wrote:

> Robert Bradshaw wrote:
>> What is the plan for P3k vs. Python 2.x? Should there be a separate
>> branch? I imagine the ParseArgumentsAndKeywords function will  
>> change to
>> reflect this in P3k, so much of this "manual labor" will be  
>> redundant.
>
> There is the "py3k" branch in Python's SVN, but since the PEP  
> doesn't say
> anything about changes to the C API, I don't expect anything to  
> happen any
> time soon. I mean, you can easily do this in plain C code already  
> by parsing
> the args accordingly. It's just not currently available in Pyrex/ 
> Cython, as
> you can't adapt the way it parses the arguments from Pyrex code.

I meant you'd be able to just pass it a string, like you do now for  
the rest of the arguments.

> The patch isn't that long, and there isn't usually that much  
> "manual labour"
> involved in the generated code. It's pretty optimised already.
>
> Regarding the general question about a 3k branch for Cython: I'd  
> rather wait
> and see if we can't handle that with conditional compilation as we  
> currently
> do with most newer features. I prefer keeping Cython a single tool  
> for all
> target architectures and versions as long as we can. And it would  
> be great to
> generate C code that compiles and works without changes for Python  
> 2.3 through
> Python 3.0. :]

This is my hope too, no idea how realistic it is though.

- Robert