2016-07-06 17:01:27 by Filip Hajny | Files touched by this commit (2) |
Log message:
Update databases/py-cassandra-driver to 3.5.0.
Features
- Optional Execution Profiles for the core driver (PYTHON-569)
- API to get the host metadata associated with the control
connection node (PYTHON-583)
- Expose CDC option in table metadata CQL (PYTHON-593)
Bug Fixes
- Clean up Asyncore socket map when fork is detected (PYTHON-577)
- cqlengine: QuerySet only() is not respected when there are
deferred fields (PYTHON-560)
|
2016-06-06 16:45:03 by Filip Hajny | Files touched by this commit (2) |
Log message:
Update databases/py-cassandra-driver to 3.4.1.
3.4.1
Bug Fixes
- Gevent connection closes on IO timeout
- "dictionary changed size during iteration" with Python 3
3.4.0
Features
- Include DSE version and workload in Host data
- Add a context manager to Cluster and Session
- Better Error Message for Unsupported Protocol Version
- Make the error message explicitly state when an error comes from
the server
- Short Circuit meta refresh on topo change if NEW_NODE already
exists
- Show warning when the wrong config is passed to SimpleStatement
- Return namedtuple result pairs from execute_concurrent
- BatchStatement should enforce batch size limit in a better way
- Validate min/max request thresholds for connection pool scaling
- Handle or warn about multiple hosts with the same rpc_address
- Write docs around working with datetime and timezones
Bug Fixes
- High CPU utilization when using asyncore event loop
- Fix CQL Export for non-ASCII Identifiers
- Make stress scripts Python 2.6 compatible
- UnicodeDecodeError when unicode characters in key in BOP
- WhiteListRoundRobinPolicy should resolve hosts
- Cluster and Session do not GC after leaving scope
- Don't wait for schema agreement on ignored nodes
- Reprepare on_up with many clients causes node overload
- None inserted into host map when control connection node is
decommissioned
- weakref.ref does not accept keyword arguments
|
2016-05-06 11:41:36 by Filip Hajny | Files touched by this commit (2) |
Log message:
Update databases/py-cassandra-driver to 3.3.0.
Features
- Add an AddressTranslator interface
- New Retry Policy Decision - try next host
- Don't mark host down on timeout
- SSL hostname verification
- Add C* version to metadata or cluster objects
- Options to Disable Schema, Token Metadata Processing
- Expose listen_address of node we get ring information from
- Use A-record with multiple IPs for contact points
- Custom consistency level for populating query traces
- Normalize Server Exception Types
- Propagate exception message when DDL schema agreement fails
- Specialized exceptions for metadata refresh methods failure
Bug Fixes
- Resolve contact point hostnames to avoid duplicate hosts
- GeventConnection stalls requests when read is a multiple of the
input buffer size
- named_tuple_factory breaks with duplicate "cleaned" col names
- Connection leak if Cluster.shutdown() happens during
reconnection
- HostConnection.borrow_connection does not block when all request
ids are used
- Empty field not being handled by the NumpyProtocolHandler
|
2016-04-21 11:31:21 by Filip Hajny | Files touched by this commit (2) | |
Log message:
Update databases/py-cassandra-driver to 3.2.2.
- Fix counter save-after-no-update (PYTHON-547)
|
2016-04-19 14:41:48 by Filip Hajny | Files touched by this commit (2) | |
Log message:
Update databases/py-cassandra-driver to 3.2.1.
Fix dependencies.
Changes in 3.2.1:
- Introduced an update to allow deserializer compilation with
recently released Cython 0.24
|
2016-04-13 14:07:05 by Filip Hajny | Files touched by this commit (2) |
Log message:
Update databases/py-cassandra-driver to 3.2.0.
Features
- cqlengine: Warn on sync_schema type mismatch
- cqlengine: Automatically defer fields with the '=' operator
(and immutable values) in select queries
- cqlengine: support non-equal conditions for LWT
- cqlengine: sync_table should validate the primary key composition
- cqlengine: token-aware routing for mapper statements
Bug Fixes
- Deleting a column in a lightweight transaction raises
a SyntaxException
- cqlengine: make Token function works with named tables/columns
- comparing models with datetime fields fail #79
- cython date deserializer integer math should be aligned
with CPython
- db_field is not always respected with UpdateStatement
- Sync_table fails on column.Set with secondary index
|
2016-04-08 18:43:39 by Filip Hajny | Files touched by this commit (2) |
Log message:
Update databases/py-cassandra-driver to 3.1.1.
- cqlengine: Fix performance issue related to additional "COUNT"
queries (PYTHON-522)
|
2016-03-12 10:53:54 by Filip Hajny | Files touched by this commit (2) | |
Log message:
Update databases/py-cassandra-driver to 3.1.0.
Features
- Pass name of server auth class to AuthProvider
- Surface schema agreed flag for DDL statements
- Automatically convert float and int to Decimal on serialization
- Eventlet Reactor IO improvement
- Make pure Python ProtocolHandler available even when Cython is
present
- Optional Cython deserializer for bytes as bytearray
- Add Session.default_serial_consistency_level
- cqlengine: Expose prior state information via cqlengine
LWTException
- cqlengine: Collection datatype "contains" operators support
(Cassandra 2.1)
- cqlengine: Add DISTINCT query operator
- cqlengine: Tuple cqlengine api
- cqlengine: Add support for UPDATE/DELETE ... IF EXISTS statements
- cqlengine: Allow nested container types
- cqlengine: Add ability to set query's fetch_size and limit
- cqlengine: Internalize default keyspace from successive set_session
- cqlengine: Warn when Model.create() on Counters
Bug Fixes
- Bus error (alignment issues) when running cython on some ARM
platforms
- Overflow when decoding large collections (cython)
- Timer heap comparison issue with Python 3
- Cython deserializer date overflow at 2^31 - 1
- Decode error encountered when cython deserializing large map
results
- Don't require Cython for build if compiler or Python header
not present
- Unorderable types in task scheduling with Python 3
- cqlengine: Fix crash when updating a UDT column with a None value
- cqlengine: Race condition in ..connection.execute with lazy_connect
- cqlengine: doesn't support case sensitive column family names
- cqlengine: UserDefinedType mandatory in create or update
- cqlengine: db_field breaks UserType
- cqlengine: UDT badly quoted
- cqlengine: Use of db_field on primary key prevents querying except
while tracing.
- cqlengine: DateType.deserialize being called with one argument
vs two
- cqlengine: Querying without setting up connection now throws
AttributeError and not CQLEngineException
- cqlengine: BatchQuery multiple time executing execute statements.
- cqlengine: Better error for management functions when no connection
set
- cqlengine: Handle None values for UDT attributes in cqlengine
- cqlengine: Fix inserting None for model save
- cqlengine: EQ doesn't map to a QueryOperator (setup race condition)
- cqlengine: class.MultipleObjectsReturned has DoesNotExist as base
class
- cqlengine: Typo in cqlengine UserType __len__ breaks attribute
assignment
Other
- cqlengine: a major improvement on queryset has been introduced.
It is a lot more efficient to iterate large datasets: the rows
are now fetched on demand using the driver pagination.
- cqlengine: the queryset len() and count() behaviors have changed.
It now executes a "SELECT COUNT(*)" of the query rather than
returning the size of the internal result_cache (loaded rows).
On large queryset, you might want to avoid using them due to
the performance cost. Note that trying to access objects using
list index/slicing with negative indices also requires a count
to be executed.
|
2015-11-25 16:58:10 by Filip Hajny | Files touched by this commit (3) | |
Log message:
Update databases/py-cassandra-driver to 3.0.0.
3.0.0
Features
- Support datetime.date objects as a DateType
- Add Cluster.update_view_metadata
- QueryTrace option to populate partial trace sessions
- Attach column names to ResultSet
- Change default consistency level to LOCAL_ONE
Bug Fixes
- Properly SerDes nested collections when protocol_version < 3
- Evict UDTs from UserType cache on change
- Make sure query strings are always encoded UTF-8
- Track previous value of columns at instantiation in CQLengine
- UDT CQL encoding does not work for unicode values
- NetworkTopologyStrategy#make_token_replica_map does not account
for multiple racks in a DC
- Cython integer overflow on decimal type deserialization
- Query trace: if session hasn't been logged, query trace can
throw exception
3.0.0rc1
Features
- Process Modernized Schema Tables for Cassandra 3.0
- Remove deprecated features
- Don't assign trace data to Statements
- Normalize results return
- Process Materialized View Metadata/Events
- Remove blist as soft dependency
- Change default consistency level to LOCAL_QUORUM
- Normalize CQL query/export in metadata model
Bug Fixes
- Implementation of named arguments bind is non-pythonic
- CQL encoding is incorrect for NaN and Infinity floats
- Protocol downgrade issue with C* 2.0.x, 2.1.x, and python3,
with non-default logging
- ValueError when accessing usertype with non-alphanumeric
field names
- NumpyProtocolHandler does not play well with PagedResult
2.7.2
Bug Fixes
- Resolve CQL export error for UDF with zero parameters
- Remove futures dep. for Python 3
|
2015-11-03 02:56:36 by Alistair G. Crooks | Files touched by this commit (368) |
Log message:
Add SHA512 digests for distfiles for databases category
Problems found with existing distfiles:
distfiles/D6.data.ros.gz
distfiles/cstore0.2.tar.gz
distfiles/data4.tar.gz
distfiles/sphinx-2.2.7-release.tar.gz
No changes made to the cstore or mariadb55-client distinfo files.
Otherwise, existing SHA1 digests verified and found to be the same on
the machine holding the existing distfiles (morden). All existing
SHA1 digests retained for now as an audit trail.
|