summaryrefslogtreecommitdiff
path: root/redis
Commit message (Collapse)AuthorAgeFilesLines
* version bump!2.4Andy McCurdy2011-06-011-1/+1
|
* pipelines need the response_callback love, too. grammar fixes in docsAndy McCurdy2011-06-011-2/+8
|
* added ability for per-instance response callbacks. also, holy shit, ↵Andy McCurdy2011-06-011-4/+17
| | | | documentation!
* updating a few docstringsAndy McCurdy2011-05-311-2/+7
|
* support ZADD variable length args. deprecate the version passing score and ↵Andy McCurdy2011-05-311-4/+18
| | | | value (in the wrong order).
* ConnectionPool's get_connection() now can take optional kwargsAndy McCurdy2011-05-252-3/+4
|
* remove the SELECT command completely. use a separate client instance instead.Andy McCurdy2011-05-241-10/+0
|
* renamed path arg to unix_socket_path to make it clearerAndy McCurdy2011-05-231-3/+3
|
* ability to use the UnixDomainSocketConnection directly from the Redis client.Andy McCurdy2011-05-231-14/+23
|
* added Connection and UnixDomainSocketConnection to module importsAndy McCurdy2011-05-231-3/+7
|
* new pubsub testsAndy McCurdy2011-05-232-2/+2
|
* removing unnecessary importsAndy McCurdy2011-05-221-4/+2
|
* removal of more deprecated commandsAndy McCurdy2011-05-221-51/+1
|
* allow shard hints to be passed to pipeline and pubsub objects. a smart ↵Andy McCurdy2011-05-221-90/+53
| | | | | | connection pool could use these hints to determine the correct shard to run on. removed all previously deprecated parameters and commands
* completely refactored pubsub. this is backwards incompatible, but quite ↵Andy McCurdy2011-05-221-70/+96
| | | | necessary.
* more error checking for socket errorsAndy McCurdy2011-05-221-5/+7
|
* connection pool is now a real pool. we no longer rely on threading.local for ↵Andy McCurdy2011-05-172-26/+31
| | | | anything in redis-py. yaa!
* all tests pass now except pub/sub. connection_pool's get_connection now ↵Andy McCurdy2011-05-172-19/+18
| | | | always received the command name for the next command. still need to pass keys.
* all tests passing with new connection poolAndy McCurdy2011-05-162-140/+138
|
* make a formal hook so other types of connections don't have to repeat the ↵Andy McCurdy2011-05-121-4/+4
| | | | same connection setup logic
* we don't need the command_name overriding just yetAndy McCurdy2011-05-121-4/+3
|
* connection class completely refactored. encoding and command packing moved ↵Andy McCurdy2011-05-122-171/+181
| | | | from client to connection. introduced concept of protocol parsers and implemented both a PythonParse and a hiredis parser. the parser class can be overridden in the __init__ of the connection if desired.
* Merged pull request #132 from banjiewen/banjiewen.Andy McCurdy2011-04-261-0/+5
|\ | | | | Added socket disconnection if commands are rejected due to AOF/RDB load.
| * Added socket disconnection if commands are rejected due to AOF/RDB load.Benjamin Anderson2011-04-221-0/+5
| | | | | | | | SELECT commands are issued only on initial socket connection. If a user issues a command during an AOF/RDB load, the socket connection will be successfully opened, but the SELECT command will be lost. If subsequent commands are sent after the AOF/RDB load is completed, the SELECT command will not be retried (due to the open socket), and those commands will go to the wrong DB. This behavior appears to be present in hiredis as well as redis-py. Fixing it here by special casing the AOF/RDB load message and closing the socket connection before raising ResponseError.
* | __getitem__ now raises a KeyError when a key doesn't existAndy McCurdy2011-04-261-3/+3
| |
* | trying to getitem on a key that doesn't exist should raise a KeyErrorIonuț Arțăriși2011-04-051-2/+12
|/
* Use IPPROTO_TCP instead of SOL_TCPAdam Vandenberg2011-03-271-6/+1
| | | | | The constant value is the same, 6, but the former works in Jython as well as CPython.
* Jython doesn't define socket.SOL_TCP, but seems to work by using ↵Andy McCurdy2011-03-271-1/+7
| | | | socket.SOL_TCP's constant value, 6. Fix for #97
* version bumpAndy McCurdy2011-03-171-1/+1
|
* slight performance tweak to zset_score_pairs. Thanks SuorAndy McCurdy2011-03-161-2/+2
|
* Some speed optimizations. Thanks Suor.Andy McCurdy2011-03-151-8/+6
|
* Always clear the subscription status of the client after a disconnect. Fixes #46Andy McCurdy2011-03-141-0/+1
|
* add PubSubError to better report disallowed command when using pub/subDan Colish2011-03-143-6/+12
|
* Add exception message for DataError being thrown in hmsetDan Colish2011-03-141-1/+1
|
* Raise DataError on empty hmset mappings, remove InvalidDataDan Colish2011-03-143-7/+23
|
* Merge branch 'andy'Dan Colish2011-03-141-2/+2
|\
| * Fix for #103, handling socket errors betterAndy McCurdy2011-03-141-2/+2
| |
* | Use itertools izip and islice to speed up pairs_to_dict and other zipsDan Colish2011-03-141-4/+6
|/
* Merge branch 'andy'Dan Colish2011-03-141-3/+3
|\
| * Merge remote branch 'richid/multi_watch' into richidAndy McCurdy2011-03-141-3/+3
| |\
| | * Add support for multiple keys in a single WATCH callRich Schumacher2011-03-041-3/+3
| | |
* | | Remove ternary operator to support Python 2.4Dan Colish2011-03-141-1/+4
|/ /
* | Lock's now support timeouts that are floats. Fix for #113Andy McCurdy2011-03-141-4/+10
| |
* | the actual order of the zrevrangebyscore parameters were switched, too, ↵Andy McCurdy2011-03-141-6/+6
| | | | | | | | which was the root of this problem
* | Really compatible with the new info in redis 2.2 : ignore blank and comment ↵twidi2011-02-271-5/+6
| | | | | | | | lines
* | parameters for zrevrangebyscore are `max` then `min`, and not `min` then `max`twidi2011-02-271-1/+1
|/
* Merge remote branch 'bretthoerner/master' into bretthoernerAndy McCurdy2011-02-091-2/+2
|\
| * Fix info() for new Redis INFO format (backwards compatible).Brett Hoerner2011-01-241-2/+2
| |
* | stylistic modes + tests for zrevrangebyscoreAndy McCurdy2011-02-091-8/+12
| |
* | Merge remote branch 'jaybaird/master' into jaybairdAndy McCurdy2011-02-091-1/+22
|\ \
View Lorry Controller Status