[Python-Dev] Proposal: explicitly disallow function/class mismatches in accelerator modules

[Nick Coghlan]

On 11 July 2016 at 13:26, Steven D'Aprano <steve at pearwood.info> wrote:
> My two core questions are:
>
> (1) How much extra effort are we going to *mandate* that core devs put
> in to hide the differences between C and Python code, for the benefit of
> a small minority that will notice them?
>
> (2) When should that effort be done? Upfront, or when and as problems
> are reported or noticed?
>
> My preference for answers will be, (1) not much, and (2) when problems
> are reported. In other words, close to the status quo.

I think it's still OK to defer fixing discrepancies between the
alternate implementations until people explicitly report a problem,
which I guess means the policy update that I'd be looking for is that
in cases where the C version currently implements a superset of the
Python version's behaviour, the recommended policy be to accept
patches to align the two even if:

- it makes the Python version more complex and hence harder to read
- it makes the Python version slower on CPython (and likely other
runtimes without even a method JIT)

Replacing a closure with a custom callable (as with the
functools.partial patch) is a case where both of those objections
legitimately apply, hence making the right thing to do less than
obvious given the current wording of PEP 399 - it isn't clear whether
those simplicity focused considerations should trump the cross-runtime
compatibility ones.

Cheers,
Nick.

-- 
Nick Coghlan   |   ncoghlan at gmail.com   |   Brisbane, Australia