(implementation/tests contributed by Sam Ruby)

of this test.  It probably still requires more disk space
than most buildbots have, and in any case is still so
intrusive that if we don't find another way to test this I'm
taking my buildbot offline permanently ;-)

This will hopefully get the buildbots to pass.  Not sure this
test will be feasible or even work.  But everything is red now,
so it can't get much worse.

  now works reliably.  It has been updated to use automatic BerkeleyDB
  deadlock detection and the bsddb.dbutils.DeadlockWrap wrapper to retry
  database calls that would previously deadlock. [SF python bug #775414]

until Gerhard has time to fully debug the issue.  This affects versions
before 3.2.1 (possibly only versions earlier than 3.1.3).

Based on discussion on python-checkins.

This backs out the test changes in 46962 which prevented crashes
by not running the tests via a version check.  All the version checks
added in that rev were removed from the tests.

Code was added to the error handler in connection.c that seems
to work with older versions of sqlite including 3.1.3.

  SQLite versions.
- Added version checks in test suite so that we don't execute tests that we
  know will fail with older (buggy) SQLite versions.

Now, all tests should run against all SQLite versions from 3.0.8 until 3.3.6
(latest one now). The sqlite3 module can be built against all these SQLite
versions and the sqlite3 module does its best to not trigger bugs in SQLite,
but using SQLite 3.3.3 or later is recommended.

mime.types file for determining MIME types.

objects of same type/of subclasses of the other.

__del__ method when initialization failed.

but that SF patch 1504676 as written mis-handles

- remove some over-zealous triple-quoting

adjacent triples in the result list describe non-adjacent matching
blocks.  That's _nice_ to have, and Guido said he wanted it.

Not a bugfix candidate:  Guido or not ;-), this changes visible
endcase semantics (note that some tests had to change), and
nothing about this was documented before.  Since it was working
as designed, and behavior was consistent with the docs, it wasn't
"a bug".