of source path to set_data() and make the new argument private until
possible API changes can be discussed more thoroughly in Python 3.4.

Patch by Chris Rebert.

Patch by Chris Rebert.

already one.

Patch by Dave Malcolm.

Closes #9374: add back now-unused module attributes; removing them is a backward compatibility issue, since they have a public-seeming name.

Patch by Vinay Sajip.

Fix by Colin Su.  Test by me, based on a test written by Petri Lehtinen.

Fix by Colin Su.  Test by me, based on a test written by Petri Lehtinen.

 Add input dependency on custom build step.

It was exposed by 03c98d05b140 and dbbf3ccf72e8.

Close #2501: Permission bits are once again correctly copied from the source file to the cached bytecode file. Test by Eric Snow.

decimal.py. The hack to remove the ideal exponent is no longer required.

environment variable to a value comparable to what it sets for CC for
the benefit of C++ extension modules.
(Patch by Ronald Oussoren)

(Patch by Ronald Oussoren)

(Patch by Ronald Oussoren)

It was correct in the original patch and I foobared it
when I restructured part of the code.

The new code correctly handles historic changes in UTC offsets.
A test for this should follow.

Original patch by Alexander Belopolsky.

had a coefficient with MAX_PREC=425000000 digits and a negative exponent.
Increasing the context limits above the official values fixes the issue
and is safe (in this case!).

   required for a corner case in dec_hash() in the following commit and also
   usually faster. dec_hash() needs some extra precision above MPD_MAX_PREC,
   and _mpd_base_ndivmod() is not audited for that.

2) Use _mpd_basemul() if the length of the smaller operand is less than
   or equal to 256. While this is technically an optimization, it is
   required for *testing* corner cases in dec_hash() in reasonable time.

This is due to an off-by-one error: the allocated buffer didn't have room for a NUL
character at the end of the mbstowcs result.

(merge with 3.2)