interfaces provided by the served storage on the storage server.

The old way this was done was to copy all of the interface
declarations from the served storage. This has 2 problems:

1. Not all interfaces copied were actually provided by the client
storage. Just because a remote storage provides an interface, doesn't
mean the client storage can, especially if the method (or attribute)
isn't supported by the ZEO protocol.

2. Older clients could get import errors while trying to import copies
interfaces.

Now, we only declare from a known set of interfaces defined on the client.

Also removed the version argument from the FileStorage history method.

history interface.

- Added a helper function, ZODB.blob.is_blob_record for testing whether
  a data record is for a blob.  This can be used when iterating over a
  storage to detect blob records so that blob data can be copied.

In the future, we may want to build this into a blob-aware
  iteration interface, so that records get blob file attributes
  automatically.

- Added the IBlobStorageRestoreable interfaces for blob storages
  that support recovery via a restoreBlob method.

- Updated ZODB.blob.BlobStorage to implement
  IBlobStorageRestoreable and to have a copyTransactionsFrom method
  that also copies blob data.

Also removed the version argument from the history method.

read-only files can't be removed on windows.

in and that the file must not be open.

windows. The test author asures me that leaving the handle open wasn't
intentional.

deal with changes in hashability of persistent lists in Python
2.6. (It's still a puzzle why they are hashable in realier versions of
Python, but not in Python 2.6.)

that is cleaned up at the start of a test run.

Unfortunately, the ZODB tests leave lots of temporary files behind
which can cause failures in subsequent test runs and which tend to
litter /tmp. Eventually, I want to clean that up, but, in the mean
time, I can limit the damage to the test part directory.

3.8 branch)

unsigned long field after a signed char state field and a 3-character
reserved field.  This didn't work because there are packages in the
wild that have their own copies of cPersistence.h that didn't see the
update.

To get around this, we used the reserved space by making
estimated_size a 24-bit bit field in the space occupied by the old
3-character reserved field.  To fit in 24 bits, we made the units of
estimated_size 64-character blocks.  This allows is to handle up to a
GB.  We should never see that, but to be paranoid, we also truncate
sizes greater than 1GB.  We also set the minimum size to 64 bytes.

  correctly, leaving the client storage in an inconsistent state.

the file storage is closed because it's internal meta data may be
invalid.

number of writes.  This was done during the last phase of two-phase
commit, which made this critical phase more subject to errors than
it should have been.  Also, for large databases, saves were done so
infrequently as to be useless.  The feature was removed to reduce
the chance for errors during the last phase of two-phase commit.