checkTransactionalUndoIterator: do not expect iterator to return sorted oids

fstail: print the txn offset and header size, instead of only the data offset

Before:

    2016-07-01 09:41:50.416574: hash=d7101c5ee7b8e412d7b6d54873204421e09b7f34
    user='' description='' length=1629 offset=58990284

After:

    2016-07-01 09:41:50.416574: hash=d7101c5ee7b8e412d7b6d54873204421e09b7f34
    user='' description='' length=1629 offset=58990261 (+23)

The structure of a FileStorage DB is such that it's easy to revert the last
transactions, by truncating the file at the right offset. With the above
change, `fstail` can now be used to get this offset.

In the above example:

    truncate -s 58990261 Data.fs

would delete the transaction and all those after.

Refactored FileStorage transactional undo

Fix handle_all_serials for the new and old protocols.

As part of a project to provide object-level commit locks for ZEO, I'm
refactiring FileStorage to maintain transaction-specific data in
Tranaction.data.  This involved undo.  In trying to figure this out, I
found:

- A bug in _undoDataInfo, which I verified with some tests and

- _transactionalUndoRecord was maddeningly difficult to reason about
  (and thus change).

I was concerned less by the bug than my inability to know whether a
change to the code would be correct.

So I refactored the code, mainly transactionalUndoRecord, to make the
code easier to understand, fixing some logic errors (I'm pretty sure)
along the way.  This included lots of comments. (Comments are much
easier to compose when you're working out logic you didn't
understand.)

In addition to makeing the code cleaner, it allows undo to be handled
in cases that weren't handled before.

Only tpc_vote can report resolved conflicts with the new commit protocol

Better support of the new API to notify of resolved conflicts (store/tpc_finish)

Conflicts:
	src/ZODB/interfaces.py

ZEO handles undo asynchronously and ZODB has for some time allowed
undo oids to be returned via vote.

Define a new interface for storages that support the new commit protocol.

When switching all storages to the new API in the master branch,
I found a few issues, mainly with blobs.

1. Created/modified blobs are invalidated during the first phase
   (in Connection._store_objects):

    obj._p_invalidate()

    And with the old API, _handle_serial then sets _p_serial
    Ghost objects are not supposed to have a _p_serial and with the new API,
    a few tests would fail because _p_serial are checked without activating the
    blob first.

2. Another consequence of _handle_serial not updating _p_changed/_p_serial
   immediately is that created objects are stored twice if __getstate__
   modifies itself. This case is tested in testConnection by
   doctest_lp485456_setattr_in_setstate_doesnt_cause_multiple_stores

    Hence the change in Connection._commit:
    - self._modified is already appended in Connection._store_objects
    - (obj._p_serial == z64) instead of (oid in self._creating)
      would not work for savepoints.

3. Setting _p_changed of a Blob with no uncommitted changes would cause an error
   (lp440234_Setting__p_changed_of_a_Blob_w_no_uncomitted_changes_is_noop)

    Fixed by the same change as in 2:
    - oid was appended twice to self._modified but reverted one
      (self._modified.pop() in _store_objects)
    - the test passed because _p_changed was reset early by _handle_serial

This allows interface incompatibilities:

- Future version of ZODB that requires the new protocol being
  configured with a storage that doesn't implement the new protocol

- ZEO has a client that uses the old protocol can't use a storage that
  implements the new protocol.

To be caught early.

This commit also clarifies that the serial returned by ``undo()`` is ignored.

This reverts commit 7901868a.

This allows interface incompatibilities:

- Future version of ZODB that requires the new protocol being
  configured with a storage that doesn't implement the new protocol

- ZEO has a client that uses the old protocol can't use a storage that
  implements the new protocol.

To be caught early.

This commit also clarifies that the serial returned by ``undo()`` is ignored.

Allow serial to be returned as late as tpc_finish

This makes possible for storage to allocate serial inside tpc_finish,
removing the requirement to serialise 2PC's second phase phase (tpc_vote
to tpc_finish/tpc_abort).
Co-Authored-By: Julien Muchembled <jm@nexedi.com>