Various edits

the results of *setting* __name__ are not so surprising.

If people can suggest more tests, that'd be grand, or is what's there
sufficient?

Convert test/README to reST

see problems with my code that I didn't see before the checkin, but:

When a subtype .mro() fails, we need to reset the type whose __bases__
are being changed, too.  Fix + test.

mro() methods.  Now any exception aborts the whole __bases__ change.

And more tests.

   of using the CCompiler.find_library_file() provided by the Distutils.
   This patch fixes it to use the Distutils method at the cost of some
   additional glue.

(The duplication resulted in the SSL module not being automatically
built on Macs; the Distutils knew that shared libraries on OS X have a
.dylib extension, but the setup.py code didn't.)

Fill in the missing details. (From Jack Jansen)

Constructor accepts optional keyword arguments after a optional items list.
Add fromkeys() as an alternate constructor from an iterable over keys.
Expand related unittests.

method for constructing new dictionaries from sequences of keys.

new-style classes.
Closes SF bug #637941.

Avoid depending on this in the test.

use regrtest, so it might be interesting.

allow the use of any resource except bsddb, give the option
"-uall,-bsddb".

  Fixed minor bug preventing documentation compilation.

It's odd that gcc on my ibook didn't complain about this.

This adds a --relative option to the bdist_dumb command that defaults
to false; if true, the .tar.gz or .zip will be assembled using relative
paths.

[ 635933 ] make some type attrs writable

but should have been!

News about the above.

[ 635933 ] make some type attrs writable

Plus a couple of extra tests beyond what's up there.

It hasn't been as carefully reviewed as it perhaps should, so all readers
are encouraged, nay exhorted, to give this a close reading.

There are still a couple of oddities related to assigning to __name__,
but I intend to solicit python-dev's opinions on these.

Mark writes in private email:

  "Modules listed in the registry was a dumb idea.  This whole scheme
  can die.  AFAIK, no one in the world uses it (including win32all
  since the last build)."

(See also SF #643711)

- requires modulefinder.py to be on sys.path
- does *not* work for Python.framework (yet), only for static builds

messages about MRO conflicts.  (Tweaks here: don't print the message,
but compare it with an expected string.)

messages about MRO conflicts.  (The tweaks include correcting spelling
errors, some refactoring to get the name of classic classes, and a
style nit or two.)

non-true values to leak in.  This threw-off equality testing which depends
on the underlying dictionaries having both the same keys and values.