scoder

@0dminnimda, could you please try to bring this into a mergeable state? It doesn't have to be complete, any improvement will help. We can add to it later, any time.

Can't say if the tests will succeed, but in any case, I suggest (for the `run` test) to - change the overloaded C++ functions to return some value instead of...

I just fixed a hash related bug while refactoring the code generation of `Dataclass.py` in https://github.com/cython/cython/commit/9cb28fb273b9bb9b2c08d4f6131e08f5a92f951c. Seems worth retrying if that still gives the same results. The issue was in...

> I think it's just a consequence that CPython puts everything in a tuple and hashes that while Cython doesn't Yes, that's most likely the reason. Tuples have a specific...

> And (where possible) use annotations over `@cython.declare` - right? Absolutely. The goal is to use the least intrusive and most readable way to get the same effect as with...

`.pxd` files use Cython syntax. They are a subset of the language, basically. I think it's enough to describe this possibility in one place (that's currently the `pure` page) and...

I suggest using this issue for coordination and PRs (any size) for the changes.

> The codeblocks I did find was in files though, so this will essentially double most of the files in the documentation. That suggests to me that we should have...

We allow inline functions in pxd files, and the main difference between function declarations and (inline) function implementations is that the latter have a function body. I would like to...

Wow, thanks. That's … huge. :) Do I understand correctly that this duplicates the C code, using two separate code writers? Or is the duplication limited to certain sections or...