Backported to release-maint23

The find_all_submodules() method in modulefinder only
looks for *.py, *.pyc, and *.pyo files.  Python
extension modules are only found if they are referenced
in import statements somewhere.

This patch uses the actual list from imp.get_suffixes().

Backported myself.

for this function has always claimed that was true, but it wasn't
verified before.  For the latest batch of "double deallocation" bugs
(stemming from weakref callbacks invoked by way of subtype_dealloc),
this assert would have triggered (instead of waiting for
_Py_ForgetReference to die with a segfault later).

this change, I think subtype_dealloc is actually a smidgen less obscure
than it was in 2.3 -- we got rid of a negation in an "if" <wink>.

expanded the test case with a piece that needs the more-complete fix.

I'll backport this to 2.3 maint.

subtype_dealloc():  This left the dying object exposed to gc, so that
if cyclic gc triggered during the weakref callback, gc tried to delete
the dying object a second time.  That's a disaster.  subtype_dealloc()
had a (I hope!) unique problem here, as every normal dealloc routine
untracks the object (from gc) before fiddling with weakrefs etc.  But
subtype_dealloc has obscure technical reasons for re-registering the
dying object with gc (already explained in a large comment block at
the bottom of the function).

The fix amounts to simply refraining from reregistering the dying object
with gc until after the weakref callback (if any) has been called.

This is a critical bug (hard to predict, and causes seemingly random
memory corruption when it occurs).  I'll backport it to 2.3 later.

HAVE_MEMMOVE define.  Just defined it on the cmdline (it doesn't
include Python.h, and Fred doesn't want to change the code).

Formerly, underlying queue was implemented in terms of two lists.  The
new queue is a series of singly-linked fixed length lists.

The new implementation runs much faster, supports multi-way tees, and
allows tees of tees without additional memory costs.

The root ideas for this structure were contributed by Andrew Koenig
and Guido van Rossum.

the file is a symlink.  Instead, use os.lstat directly, if it exists;
fall back on os.stat or the built-in open.  Thanks to Iustin Pop.

(same as commit of Sun Nov 2 to the release23-maint branch)

(same as commit of Sun Nov 2 to the release23-maint branch)

(same as commit of Sun Nov 2 to the release23-maint branch)

added a specific \ref to 5.9 in lieu of previous vague "see above".
(same as commit of Sun Nov 2 to the release23-maint branch)

sf bug#812818.
(same as commit of Sun Nov 2 to the release23-maint branch)

(same as commit of Sun Nov 2 to the release23-maint branch)

is exhausted.