Need to make sure that preprocessor directives start in first column.
This means we can't indent code which has preprocessor directives,
nor have a space between [ #include for example.

packages can check that extension modules are built for the right type
of python. Current values can be static, framework, shared and cfm (for
completeness, for MacPyton-OS9).

Closes bug #691889. The reporter suggests backporting this to 2.2.3
and I think I agree.

On a 64-bit machine, a dictionary could contain duplicate int/long keys
if the value was > 2**32.

a serious slowdown when loading dynamic modules that depend on large shared
libraries or frameworks.

Don't bother testing os.getlogin() if we aren't running from a tty (terminal)
It fails when run without a tty (e.g., when run from cron).

(patch provided by Greg Chapman)

Also use True/False instead of 1/0 for symlink flag.

Subsumed times() into repeat().
Added cycle() and chain().

called to find tb_lineno -- even if Py_OptimizeFlag is true.

So don't call it again when printing the traceback.

This still falls back to helpers in copy_reg for:
   - pickle protocols < 2
   - calculating the list of slot names (done only once per class)
   - the __newobj__ function (which is used as a token but never called)

test_nonrecursive_deep():  Reduced nesting depth to 60.

Not a bugfix candidate.  2.3 increased the number of stack frames
needed to pickle a list (in order to get implement the "list
batching" unpickling memory optimization new in 2.3).

seconds making the doc for get_completer marginally less minimal :)

is required for the chosen internal encoding in the init function,
as this seems to have a better chance of working under Irix and
Solaris.

Also change the test character from '\x01' to '0'.

This might fix SF bug #690309.

on the boxes I use.

is a window manager and we can connect to it, i.e. if it is safe to try
and put up windows.

As a side effect the first call will make the current process frontmost.

between str, unicode, UserString and the string module
as possible. This increases code coverage in stringobject.c
from 83% to 86% and should help keep the string classes
in sync in the future. From SF patch #662807

time.sleep(1) sometimes delays for fractionally less than a second
resulting in too short of an interval for C's time.time() function
to create a distinct seed.

the PyInt_AsLong function, and this returns a long, the value is first
retrieved with PyLong_AsLong, but afterwards overwritten by a call to
PyInt_AS_LONG.

Fixes SF #690253.

called something else!).  I can't imagine removing the prototype is
going to hurt, but put it back if *you* can.

Use level=3 (i.e. log.WARN) for the warnings about failed imports.
(Hmm...  Why is that code in an "if 1: ..."?  What's the else branch
for?)