* Fixed: A blob misfeature set blob permissions so that blobs and blob
  directories were only readable by the database process owner, rather
  than honoring user-controlled permissions (e.g. umask).

Tests running and passing on windows.

Fix PyPI page

Update RelStorage information in multi-zodb-gc document

FileStorage: Report problem on read-only open of non-existent file

... instead of silently creating empty database on such opens.

Use-case for this are utilities like e.g. zodbdump and zodbcmp which
expect such storage opens to fail so that the tool can know there is no
such storage and report it to user.

In contrast current state is: read-only opens get created-on-the-fly
empty storage with no content, but which can be iterated over without
getting any error.

This way e.g. `zodbdump non-existent.fs` produces empty output _and_
exit code 0 which is not what caller expects.

Add info on Newt DB

RelStorage 2.0 can and should use zc.zodbdgc for reference collecting and packing in a multi-db, so long as its history free.

For the record, I udes `zest.releaser` `longtest` script to debug...

TransactionUndoStorage tests close the conn and db

Notify storages that transactions have completed.

Where it makes more sense. Hysterically, it didn't matter because one
was an alias for the other.

For safety, closing the database aborts transaction managers
associated with open connections.  In explicit mode, this can lead to
NoTransaction errors.

It's likely a bug for DB objects to abort transactons on
close.  This needs more thought.

  resources after transaction complete.  See:
  https://github.com/zodb/relstorage/issues/147

- Take advantage of the new transaction-manager explicit mode to avoid
  starting transactions unnecessarily when transactions end.

Don't manipulate Connection state after it has been returned to the pool.

Good catch. Thanks.

Doing so leads to race conditions.

In particular, there can be an AttributeError.

See https://github.com/zodb/zodbshootout/issues/26 for details.

Don't require persistent at setup time. We don't build native code th…

Don't require persistent at setup time. We don't build native code that needs those headers anymore. Fixes #119.

Connection.new_oid delegates to its storage, not the DB.

This fixes leaking connections (relying on the GC to close them).

Tested with RelStorage.

Fixes #138

Fixes #139