to support surrogate characters

_Py_char2wchar() callers usually need the result size in characters. Since it's
trivial to compute it in _Py_char2wchar() (O(1) whereas wcslen() is O(n)), add
an option to get it.

surrogate characters.

surrogate characters.

new test specific to unicode (non-ascii name and filename).

Fix my previous commit (r85569).

filenames encoded to the filesystem encoding with surrogateescape error handler
(to support undecodable bytes), instead of UTF-8 in strict mode.

It occurs if the path contains surrogates.

* Expose the build flags to Python as sys.abiflags
* Shared library libpythonX.Y<abiflags>.so
* python-config --abiflags
* Make two distutils tests that failed with --enable-shared (even before this
  patch) succeed.
* Fix a few small style issues.

Patch by Sébastien Sablé.

filesystem encoding instead of utf-8.

imp_cache_from_source() encodes the input path to filesystem encoding and this
path is passed to make_compiled_pathname().

path

module path

importing encoded modules that do not depend on specific stdlib
modules being encoded in a certain way.