Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
| * | | | | Fixed issues regarding Python3 | pfreixes | 2016-06-14 | 1 | -4/+4 | |
| | | | | | ||||||
| * | | | | Geo Commands, return Python native types | Pau Freixes | 2016-06-14 | 1 | -4/+44 | |
| | | | | | ||||||
| * | | | | Implemented support for the GEO commands for Redis 3.2.0 | pfreixes | 2016-05-16 | 1 | -0/+129 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | The following commands have been implemented GEOADD, GEODIST, GEOHASH, GEOPOS, GEORADIUS, GEORADIUSBYMEMBER. | |||||
* | | | | | Merge pull request #751 from cjsimpson/token-cache | Andy McCurdy | 2016-06-14 | 2 | -29/+48 | |
|\ \ \ \ \ | | | | | | | | | | | | | Improve performance for Python 3.x | |||||
| * | | | | | Cache Token objects to improve performance. | Chris Simpson | 2016-06-12 | 2 | -29/+48 | |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The Token class now contains a cache of tokens, and each token stores its encoded value. In Python 3 this prevents encoding the Token commands (get, set, incr, etc...) repeatly. There is also a smaller performance improvement by creating fewer objects. A very basic benchmark script was also added. | |||||
* | | | | | | pubsub: improve error reporting if caller forgets to subscribe | Greg Ward | 2016-06-12 | 1 | -0/+4 | |
|/ / / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is an easy mistake to make -- at least, I keep making it. It formerly resulted in a confusing crash, "AttributeError: 'NoneType' object has no attribute 'can_read'", from parse_response(). I have had to dig into the redis-py source code more than once to figure out what went wrong. With this patch, it still crashes, but with a clearer error that clarifies what the calling code forgot to do. Fixes issue #716. | |||||
* | | | | | Retry `select` calls on `InterruptedError` | Carlton Gibson | 2016-06-10 | 2 | -2/+12 | |
| | | | | | | | | | | | | | | | | | | | | Ref #738 | |||||
* | | | | | Merge pull request #743 from closeio/retry-recv-eintr | Andy McCurdy | 2016-06-05 | 2 | -4/+62 | |
|\ \ \ \ \ | | | | | | | | | | | | | Automatically retry EINTR for Python < 3.5 to prevent duplicate command execution | |||||
| * | | | | | Style fixes (2) | Thomas Steinacher | 2016-05-24 | 1 | -2/+2 | |
| | | | | | | ||||||
| * | | | | | Style fixes | Thomas Steinacher | 2016-05-24 | 1 | -2/+3 | |
| | | | | | | ||||||
| * | | | | | For Python < 3.5, automatically retry EINTR | Thomas Steinacher | 2016-05-24 | 2 | -4/+61 | |
| | | | | | | ||||||
* | | | | | | Check `int` also not only `long` | Jeong YunWon | 2016-05-30 | 1 | -1/+1 | |
|/ / / / / | | | | | | | | | | | | | | | | PyPy can generate an `int` instance here. | |||||
* | | | | | add daemon option to PubSubWorker so ctrl-c will work | Keith Ainsworth | 2016-05-17 | 1 | -3/+4 | |
|/ / / / | ||||||
* | | | | PEP8 fixes | Tim Savage | 2016-03-29 | 1 | -5/+11 | |
| | | | | ||||||
* | | | | Extend ConnectionPool.to_url to parse querystring arguments to correct type. | Tim Savage | 2016-03-29 | 1 | -3/+30 | |
| | | | | | | | | | | | | | | | | | | | | Previously if a value for socket_timeout was supplied as part fo the URL an error would be raised when a socket was created with an invalid type, this change fixes that by parsing `socket_timeout`, `socket_connect_timeout` to float values. In addition the boolean values `socket_keepalive` and `retry_on_timeout` are parsed to bool types taking into account the usage of True/False, Yes/No strings. | |||||
* | | | | Merge pull request #685 from iandyh/master | Andy McCurdy | 2016-02-18 | 1 | -1/+49 | |
|\ \ \ \ | | | | | | | | | | | Redis cluster commands | |||||
| * | | | | redis cluster commands | iandyh | 2015-12-01 | 1 | -1/+49 | |
| |/ / / | ||||||
* | | | | Fix typo in comment. | Greg Ward | 2015-12-09 | 1 | -2/+2 | |
|/ / / | ||||||
* | | | FIX #651 try next Sentinel node on TimeoutError | Bo Lopker | 2015-11-27 | 1 | -3/+4 | |
| | | | ||||||
* | | | fix getitem syntax for redis get when value is the empty string | Javier Candeira | 2015-11-25 | 1 | -1/+1 | |
| | | | ||||||
* | | | raise TimeoutError if a socket.timeout is raised while connecting | Andy McCurdy | 2015-11-09 | 1 | -0/+2 | |
| | | | | | | | | | | | | fixes #675 | |||||
* | | | 2.10.5 (incrementing the version to avoid a pypi issue)2.10.5 | Andy McCurdy | 2015-11-02 | 1 | -1/+1 | |
| | | | ||||||
* | | | 2.10.42.10.4 | Andy McCurdy | 2015-11-02 | 1 | -1/+1 | |
| | | | ||||||
* | | | socket errors on windows contain more than 2 arguments. | Andy McCurdy | 2015-11-02 | 1 | -3/+4 | |
| | | | | | | | | | | | | fixes #641 | |||||
* | | | typo | Andy McCurdy | 2015-11-02 | 1 | -2/+2 | |
| | | | ||||||
* | | | ignore errors raised in SocketBuffer's close method. See #633. | Andy McCurdy | 2015-11-02 | 1 | -2/+10 | |
| | | | ||||||
* | | | max_connections should default to None | Andy McCurdy | 2015-11-02 | 1 | -1/+1 | |
| | | | ||||||
* | | | Merge pull request #665 from ohadpartuck/connection_pool_max_connections | Andy McCurdy | 2015-11-02 | 1 | -2/+4 | |
|\ \ \ | | | | | | | | | adding max_connections option to StrictRedis Class | |||||
| * | | | adding max_connections option to StrictRedis Class | Ohad Perry | 2015-10-13 | 1 | -2/+4 | |
| | | | | ||||||
* | | | | cleanup | Andy McCurdy | 2015-11-02 | 1 | -7/+2 | |
| | | | | ||||||
* | | | | decreased length of exception class retrieving line | th13f | 2015-10-26 | 1 | -1/+4 | |
| | | | | ||||||
* | | | | parsing 'max number of clients reached' as ConnectionError | th13f | 2015-10-26 | 1 | -2/+9 | |
| | | | | ||||||
* | | | | -Add CHANGES message. | Christian Bodt | 2015-10-20 | 1 | -3/+3 | |
| | | | | | | | | | | | | | | | | -Rename variables for coherence | |||||
* | | | | -added zrevrangebylex | Christian Bodt | 2015-10-20 | 1 | -0/+16 | |
|/ / / | ||||||
* | | | satisfy pep8 updates | Andy McCurdy | 2015-09-28 | 1 | -13/+26 | |
| | | | ||||||
* | | | removed the proactive check in HiredisParser for a line ending | Andy McCurdy | 2015-09-28 | 1 | -9/+0 | |
| | | | | | | | | | | | | | | | it turns out just calling into hiredis to test this is faster than doing string compare in Python. fixes #615 and #650. | |||||
* | | | Avoid needlessly making temporary lists to merge dict(s) together | Joshua Harlow | 2015-09-27 | 1 | -1/+2 | |
|/ / | ||||||
* | | support multiple keys for pfcount call | Chris Wilkes | 2015-05-22 | 1 | -3/+3 | |
| | | ||||||
* | | fix: Connection.encode complains when value is an object having unicode ↵ | Eric Du | 2015-05-04 | 1 | -1/+1 | |
| | | | | | | | | characters in its printable representation | |||||
* | | Fix pep8 errors | Paul Keene | 2015-02-10 | 1 | -17/+32 | |
| | | ||||||
* | | Handle percent-encoded URLs in parsing code | Paul Keene | 2015-02-09 | 2 | -9/+26 | |
| | | ||||||
* | | Add WAIT command for Redis synchronous replication | eshizhan | 2015-01-15 | 1 | -0/+9 | |
| | | | | | | | | | | ref1: https://github.com/antirez/redis/blob/unstable/src/replication.c#L1783, at SYNCHRONOUS REPLICATION ref2: http://antirez.com/news/66 This command for Redis-3.0.0 and above now. | |||||
* | | Merge branch 'master' of github.com:andymccurdy/redis-py | Andy McCurdy | 2015-01-02 | 2 | -3/+19 | |
|\ \ | |/ | ||||||
| * | Allow delay between watch errors | Joshua Harlow | 2014-12-08 | 1 | -0/+4 | |
| | | | | | | | | | | | | | | | | | | When a watcher error occurs (due to some key being watched being mutated) the current behavior is to immediately try again. To avoid the thundering herd problem a delay is nice to provide to avoid these situations by introducing a sleep period between these types of failures. | |||||
| * | fix pep8 | Hendrik Muhs | 2014-12-04 | 1 | -1/+2 | |
| | | ||||||
| * | UnicodeDecodeErrorfix unicode encode error when using pipeline in ↵ | Hendrik Muhs | 2014-12-04 | 2 | -3/+14 | |
| | | | | | | | | combination with msgpack and lua | |||||
* | | better handling of the pub/sub workerthread's stop behavior | Andy McCurdy | 2015-01-02 | 1 | -20/+29 | |
|/ | | | | | properly unsubscribes and cleans up the connection now. also allows callbacks to call thread.stop() without errors. | |||||
* | should also reinitialize with check_connection | Andy McCurdy | 2014-09-18 | 1 | -0/+1 | |
| | ||||||
* | Passing is_master flag through on reset | Jonathan Frawley | 2014-09-18 | 1 | -0/+1 | |
| | | | | | | | - is_master is not passed through when recreating the ConnectionPool. - This was causing unexpected behaviour where we would specifiy that we wish to connect to the slave, but the master would be read from instead. | |||||
* | docstring | Andy McCurdy | 2014-08-15 | 1 | -3/+9 | |
| |