huge start and stop arguments. Add tests.

evil - if the last grubbed buffer didn't happen to be the right one,
you couldn't remedy.

Mainline emacs compat - don't use third arg to buffer-substring (which
was for explicitly identifying the buffer in which to seek the
substring, and which turns out to be unnecessary).

administrators can still not add packages.

- don't use min() and max()
- interpret negative start/stop argument like negative slice indices

struct_passwd attributes and +=.

From SF patch #755245.

Also, modified UserList.index() to match and expanded the related tests.

A change from Duncan Booth, to deal with changes in the way pgen gets
built.  Note that graminit.[ch] aren't normally built on Windows (they're
obtained from CVS).

operate on the module dictionary directly.  This prevents spurious
depreciation warnings from being raised if a submodule name shadows
a builtin name.

SF bug [ 751276 ] cPickle doesn't raise error, pickle does (recursiondepth)

Most of the calls to PyErr_Clear() were intended to catch & clear an
attribute error and try something different.  Guard all those cases
with a PyErr_ExceptionMatches() and fail if some other error
occurred.  The other error is likely a bug in the user code.

This is basically the C equivalent of changing "except:" to
"except AttributeError:"

build directory is found to exist we don't clean it up. We also
use configure -C. All this lets us keep build directories, which
graeatly speeds up the process of debugging installers.

to /. Pathnames are correct for installing to / since the DESTDIR patch.

manager (if it isn't needed).

- Build idle from new location
- use ln -fsn when linking files.

besides plain tuples.

tell Windows Truth too.

Bugfix candidate.

Add explicit text for where an obindex call for generators was since obindex does not cause output of its argument.

Return None to signal that the module the object was defined in was not found when object has no __name__ attr but is needed to figure out location of object.

unique boundary strings within a program run are guaranteed.  On Windows,
duplicates were pretty likely, due to the coarse granularity of time.time.
Toned down the absurdly optimistic claims in the docstring.

Bugfix candidate.

Also install the start script from Tools/script/idle.  (This may
change again.)

Tools/idle, in both source and destination.  (There are still problems
when running the IDLE icon, but they don't seem to have to do with the
installer.)