for the tests, and restored properly when done

possible that the callback-less flavors of the ref or proxy could have
been added during GC, so we don't want to replace them.

Will backport to 2.3.

Python was built with VC 7.1.

It's not the 100% solution -- it may not even be the 90% solution -- but
it does seem to help.

an initlog() (and initlog()'s docstring says "don't use this"!) This
allows test_cgi to be run repeatedly in a single run of the interpreter.

way to guarantee a deadlock on the next call!

If I've inadvertently done some damage to this test, sorry (but I don't
think I have).

Hard to believe it ever helped anything, and it hurts finding reference
leaks.

the case of __del__ resurrecting an object.
This makes the apparent reference leaks in test_descr go away (which I
expected) and also kills off those in test_gc (which is more surprising
but less so once you actually think about it a bit).

connect error

comma expression in listpop() that was being returned.  Still essentially
unused (as it is meant to be), but now the compiler thinks it is worth
*something* by having it incremented.

- fix broken assert statement; should just raise

have differing refcount semantics.  If anyone sees a prettier way to
acheive the same ends, then please go for it.

I think this is the first time I've ever used Py_XINCREF.

(or to whatever the 'insert' method chose to return).

must have annoyed me at some point.