- bigmemtest is replaced by precisionbigmemtest
- add a poor man's watchdog thread to print memory consumption

- bigmemtest is replaced by precisionbigmemtest
- add a poor man's watchdog thread to print memory consumption

metaclass is registered with copyreg.  Patch by Nicolas M. Thiéry and
Craig Citro.

metaclass is registered with copyreg.  Patch by Nicolas M. Thiéry and
Craig Citro.

Ensure these states with assertions in _PyUnicode_CheckConsistency().

Cleanup also the code.

state

Add also related assertions to _PyUnicode_CheckConsistency().

PyUnicode_CopyCharacters() may copies less character than requested size, if
the input string is smaller than the argument. (This is very unlikely, but who
knows!?)

Avoid also calling PyUnicode_CopyCharacters() if the string is empty.

Don't check PyUnicode_CopyCharacters() in release mode. Rename also some
variables.

Remove also the FIXME for resize_copy(): as discussed with Martin, copy the
string on resize if the string is not resizable is just fine.

doesn't always set the MSG_TRUNC flag when a truncated datagram is received.

It is is called from _PyUnicode_Dump() and so must not fail.

ascii=1 is no more reserved to PyASCIIObject. Use
PyUnicode_IS_COMPACT_ASCII(obj) to check if obj is a PyASCIIObject (as before).

Only use _PyUnicode_READY_REPLACE() on just created strings.

 * resize_compact() updates also wstr_len for non-ascii strings sharing wstr
 * resize_inplace() updates also utf8_len/wstr_len for strings sharing
   utf8/wstr

Fix also a compiler warning.