What’s new in PyPy 2.6.1

Correctly resolve the output dtype of ufunc(array, scalar) calls.

Update stdlib to version 2.7.10

The JIT no longer performs loop unrolling if the loop compiles to too much code.

Build cffi import libraries as part of translation by monkey-patching an additional task into translation

Use a compact strategy for Python lists that mix integers and floats, at least if the integers fit inside 32 bits. These lists are now stored as an array of floats, like lists that contain only floats; the difference is that integers are stored as tagged NaNs. (This should have no visible effect! After lst = [42, 42.5], the value of lst[0] is still not the float 42.0 but the integer 42.)

Part of cffi 1.2.

Part of cffi 1.2.

Partial implementation of unicode dtype and unicode scalars.

Improve compatibility with numpy dtypes; handle offsets to create unions, fix str() and repr(), allow specifying itemsize, metadata and titles, add flags, allow subclassing dtype

Refactor array indexing to support ellipses.

Allow the docstrings of built-in numpy objects to be set at run-time.

Implement nditer ‘buffered’ flag and fix some edge cases

Allow multiple axes in ufunc.reduce()

Update tinylang goals to match current rpython

Clean up of vmprof, notably to handle correctly multiple threads

Remove extra link library from Boehm GC