Thanks to Nick Coghlan for pointing out that I'd forgotten about class
decorators.

in the simple path (not correctly rounded but faster).

Underflow to zero hasn't changed: _mpd_qexp() internally uses MIN_EMIN,
so the result would also underflow to zero for all emin > MIN_EMIN.

In case digits are left, the informal argument is as follows: Underflow can
occur only once in the last multiplication of the power stage (in the Horner
stage Underflow provably cannot occur, and if Underflow occurred twice in
the power stage, the result would underflow to zero on the second occasion).

Since there is no double rounding during Underflow, the effective work
precision is now 1 <= result->digits < prec. It can be shown by a somewhat
tedious argument that abs(result - e**x) < ulp(result, result->digits).

Therefore the correct rounding loop now uses ulp(result, result->digits)
to generate the bounds for e**x in case of Underflow.

options.  Patch by Roger Serwy.

the dll importer on windows to fail.

Issue #14963: Added test cases for contextlib.ExitStack exception handling behaviour (Initial patch by Alon Horev)

PyObject_GC_Resize() with incorrect error handling.  In case of errors,
the original object would be leaked.  This checkin fixes those cases.

Or perhaps more hackish, depending on your perspective.  But at least this
way it is now possible to run the individual tests using the unittest CLI.

_elementtree, with a test.

Close #14947: add missing cross-reference to Language Definition from the new dynamic type creation functions. Also cleaned up the general wording of the docs

subprocess and trace modules.

byte literals supported in Python 3.3.

Patch by Oleg Plakhotnyuk.