summaryrefslogtreecommitdiff
path: root/redis/client.py
Commit message (Collapse)AuthorAgeFilesLines
...
* [cosmetic] stupid 80 charactersAndy McCurdy2010-04-011-3/+6
|
* Handling that socket.error's args is either (errno, "msg") or "msg"Konstantin Merenkov2010-04-021-2/+6
|
* added support for BGREWRITEAOFAndy McCurdy2010-04-011-0/+7
|
* Pipeines can not optionally be transactions (wrapped in MULTI/EXEC) or not ↵Andy McCurdy2010-03-311-18/+46
| | | | by passing the transaction parameter. This fixes #23.
* Not consuming responses for subscribe commands.Konstantin Merenkov2010-04-011-4/+2
|
* [cosmetic] a bit of formating + fixed a typoKonstantin Merenkov2010-04-011-4/+4
|
* [cosmetic] removed a nearly-dupe comment from _execute_commandKonstantin Merenkov2010-04-011-2/+1
|
* Move towards binary-safeness of keys and values.Konstantin Merenkov2010-04-011-126/+119
| | | | | | | | | * Dropped support of all protocols except multi-bulk the only protocol that yet to be released redis 2.0 is able understand * As a side-effect there is an 'execute_command' function exposed to the library user. As for now it is pretty useless :) * Added (only) two tests
* [cosmetic] whitespaces cleanupKonstantin Merenkov2010-04-011-169/+172
|
* Prevent crash when using pipelines (added self.subscribed to __init__)Konstantin Merenkov2010-03-311-0/+1
|
* Yield tuple (type, channel, message) from listen() instead of dictionaryKonstantin Merenkov2010-03-311-3/+3
|
* added support for subscribe, unsubscribe, publish. tests incoming. pattern is:Andy McCurdy2010-03-301-2/+46
| | | | | | | | | redis.subscribe('my_channel') for msg in redis.listen(): channel, data = msg['channel'], msg['message'] ... listen() is a generator that will generate new messages as they come in. it will continue to block until all channels are unsubscribed.
* * fix for #21, make only the "db" argument of select() required. If host ↵Andy McCurdy2010-03-261-4/+23
| | | | | | and port aren't specified, the host/port options on the existing connection are used. * added HINCRBY support
* HEXISTS and HLEN implementationsAndy McCurdy2010-03-181-12/+26
| | | | added a socket_timeout parameter to allow commands to timeout and raise an error
* fix for #18 -- i had copy/pasted some code from another function and forgot ↵Andy McCurdy2010-03-161-1/+1
| | | | to actually change it! thanks to Sam (SJD) for the bug report
* HKEYS HVALS and HGETALL implementedAndy McCurdy2010-03-161-0/+16
|
* HDEL implementedAndy McCurdy2010-03-151-2/+6
|
* added ZRANK and ZREVRANK commands.Andy McCurdy2010-03-151-2/+17
|
* no reason to join the names here -- just let the format_inline do it, that ↵Andy McCurdy2010-03-121-1/+1
| | | | way we get the benefit of encoding, too!
* Fixed a bug with pipeline execution of AUTH and SELECT commands. These ↵Andy McCurdy2010-03-121-2/+2
| | | | | | failed after MULTI and EXEC were added, since you can't run MULTI until you've auth'd, and it doesn't make sense to run SELECT within a multi/exec. Thanks Pat Shields for the bug report
* Flushed out the connection manager object a bit more.Andy McCurdy2010-03-121-8/+22
|
* fixed #15 where zscore previously failed for members of a sorted set that ↵Andy McCurdy2010-03-111-1/+2
| | | | | | didn't exist. also made anal cleanup changes removing whitespace before colons!
* all commands being executed in a pipeline are now executing with a single ↵Andy McCurdy2010-03-111-2/+3
| | | | socket send call. this should increase network throughput
* Pipeline objects are now executed atomically via the MULTI and EXEC commandsAndy McCurdy2010-03-101-13/+57
|
* HSET now returns 1 if it added a new field to the hash, otherwise 0.Andy McCurdy2010-03-101-2/+4
| | | | There was no test for RPUSH, now there is :)
* HSET now always returned True, which also fixed part of the testAndy McCurdy2010-03-101-0/+1
|
* added HGET/HSET commandsAndy McCurdy2010-03-101-1/+10
| | | | removed the KEYS callback -- 1.34 Redis servers now return KEYS with the multi-bulk protocol, which means they're already in a list
* the Redis class is now a thread local. This means you can safely pass it ↵Andy McCurdy2010-03-051-1/+1
| | | | around between threads and connections will get swapped out correctly.
* Merge remote branch 'tabo/master'Andy McCurdy2010-03-041-2/+6
|\
| * adding support for the new lpush/rpush returning the list lengthGustavo Picon2010-03-041-2/+6
| |
* | Merge branch 'master' of github.com:andymccurdy/redis-pyAndy McCurdy2010-03-041-10/+11
|\ \ | |/
| * docstring typosZak Johnson2010-03-031-9/+9
| |
| * Add Redis.__contains__ to support "'key' in redis"Zak Johnson2010-03-031-1/+2
| |
* | complete analness about the :Andy McCurdy2010-03-041-10/+10
|/
* zrange/zrevrange/zrangebyscore should always return None if the key is ↵andymccurdy2010-02-191-1/+1
| | | | insvalid, even if withscores=True is passed
* key lists passed to blpop and brpop are not unmodified. thanks to Patrick ↵andymccurdy2010-02-181-0/+2
| | | | Shields for finding this bug
* removing the decode() hook from earlier. it's really more effective to ↵andymccurdy2010-02-171-6/+0
| | | | serialize/deserialize the data at a higher level.
* Redis.pop() is deprecated like Redis.push().andymccurdy2010-02-171-0/+3
| | | | Added a number of tests for the SORT command
* fixed a pipeline bug where a connection that had timed out and needed to ↵andymccurdy2010-02-161-6/+10
| | | | re-auth or re-select the appropriate database would end up adding the auth/select command to the end of the pipeline rather than executing immediately.
* added accessors for host/port/db information off the current connectionandymccurdy2010-02-161-0/+13
|
* - added ability for subclasses of Redis to provide their own decode logic if ↵andymccurdy2010-02-161-0/+6
| | | | | | desired. - setup.py needed to specify the package rather than py_module in order for build to work correctly.
* - connection objects now store authentication credentials so that reconnect ↵andymccurdy2010-02-161-14/+31
| | | | | | | works as expected. - Connection.connect now calls back into the redis instance class to send the auth and select commands to the server. this does not yet work with pipelines, which will be addressed shortly. - Added very basic tests for the connection pool and pipeline
* new client passes all command tests, still need tests for connection pool ↵andymccurdy2010-02-151-0/+885
and pipelining.