Antoine Pitrou wrote:
PEP 309 was written, discussed, approved, and implemented - that's how partial ended up in the stdlib.
Ok, I'm surprised that a single addition to a module needed a PEP in order to be approved.
It makes a little more sense once you realise that there was no functools module before the implementation of PEP 309. The other functions it contains in Python 2.5 (update_wrapper() and wraps()) were added later in the development cycle and reduce() didn't get added to it until 2.6/3.0.
If a concrete proposal is made that emphasises the improved introspection capabilities and raw speed increase that a function composition approach can offer over the use of lambda then I'd personally be willing to support this idea, since it was at least in part those two ideas that sold the idea of partial(). (partial() did have a big advantage over compose() in that the former's readability gains were far more obvious to most readers).
Cheers, Nick.
P.S. PEP 309 is wrong when it says a C version probably isn't worthwhile - between the time the PEP was first implemented and the time 2.5 was actually released, enough investigation was done to show that the speed gain from Hye-Shik Chang's C version was well worth the additional code complexity.