(backport from rev. 54939)

Add code to read from master_fd in the parent, breaking when we get an OSError
(EIO can occur on Linux) or there's no more data to read.  Without this,
test_pty.py can hang on the waitpid() because the child is blocking on the
stdout write.  This will definitely happen on Mac OS X and could potentially
happen on other platforms.  See the comment for details.

not cleared before __del__ is run.
 (backport from rev. 54918)

Removed obsolete comment, since .dll modules are no longer supported on windows, only .pyd.

neither documented nor intended but works in Python 2.4 under certain
conditions. This stopped working in 2.5 because struct.pack is used.
This small patch restores the old behaviour. A more solid solution is
planned for 2.6.

Fix utf-8-sig incremental decoder, which didn't recognise a BOM when the
first chunk fed to the decoder started with a BOM, but was longer than 3 bytes.

argument.

A regression from rev. 52525.

keys() if subclassed.  This is to remain consistent with 2.5.

See discussion here:
  http://mail.python.org/pipermail/python-dev/2007-April/072565.html

Revert bogus asserts (added to the wrong place) from rev 52501.

Approved by Anthony.

Fix a bug when using the __lltrace__ opcode tracer, and a problem sith signed chars in frameobject.c which can occur with opcodes > 127

(EIO can occur on Linux) or there's no more data to read.  Without this,
test_pty.py can hang on the waitpid() because the child is blocking on the
stdout write.  This will definitely happen on Mac OS X and could potentially
happen on other platforms.  See the comment for details.

  Patch taken from http://bugs.debian.org/416934.