summaryrefslogtreecommitdiff
path: root/zake
Commit message (Expand)AuthorAgeFilesLines
* Changed attribute name "_child_watches" to "_child_watchers".Kenneth Myers2015-03-021-6/+6
* Ensure kill command is sent as byte stringJoshua Harlow2015-02-131-1/+1
* Result from command() should be regular string, not bytesJoshua Harlow2015-02-131-1/+1
* Ensure commands are always byte stringsJoshua Harlow2015-02-132-8/+8
* Add basic envi command supportJoshua Harlow2015-02-132-0/+14
* Changed attribute name "_data_watches" to "_data_watchers".Kenneth Myers2015-01-151-7/+7
* Unify the exception messaging for no node foundJoshua Harlow2014-09-161-1/+1
* Correctly handle get_children() with a path that does not existJoshua Harlow2014-09-122-3/+16
* Copy the incoming sequences if providedJoshua Harlow2014-08-311-1/+1
* Expose the storage properties directlyJoshua Harlow2014-08-311-9/+8
* Just take the wait method from 2.7Joshua Harlow2014-08-311-8/+8
* Fix how the custom event is only needed on <= 2.6Joshua Harlow2014-08-311-8/+13
* Fix the overline lengthJoshua Harlow2014-08-311-1/+2
* Use a custom event and use a timeout to avoid waiting foreverJoshua Harlow2014-08-311-22/+35
* Clear the watches in the stop methodJoshua Harlow2014-08-311-0/+3
* Do the watch connected checking in the fire methodJoshua Harlow2014-08-311-8/+7
* More better variable namesJoshua Harlow2014-08-311-12/+12
* Put the path deletions into there own mini-transactionJoshua Harlow2014-08-311-18/+19
* Remove extra newlineJoshua Harlow2014-08-311-1/+0
* Decided to revert that last change and keep the existing logicJoshua Harlow2014-08-311-24/+20
* Avoid holding the open_close_lock while performing activitiesJoshua Harlow2014-08-311-12/+18
* Avoid creating a secondary list duplicateJoshua Harlow2014-08-311-1/+1
* Just use a signal watch mutation lockJoshua Harlow2014-08-311-13/+9
* Use the more natural int() operator on the uuidJoshua Harlow2014-08-311-1/+1
* Ensure all test threads are made as daemon threadsJoshua Harlow2014-08-311-8/+12
* After further reading locks seem needed around mutating operationsJoshua Harlow2014-08-311-1/+2
* More use better named variablesJoshua Harlow2014-08-311-3/+3
* Use better named variablesJoshua Harlow2014-08-301-3/+4
* Use a more meaningful variable nameJoshua Harlow2014-08-301-4/+6
* Always dispatch for all paths (instead of skipping)Joshua Harlow2014-08-301-4/+0
* Pop and contains also don't need to use the lockJoshua Harlow2014-08-301-4/+2
* Avoid requiring the lock on attempts to pop the root pathJoshua Harlow2014-08-301-3/+3
* Separate the path gathering from watch creationJoshua Harlow2014-08-301-9/+10
* Just use the clients property which already locks itselfJoshua Harlow2014-08-301-3/+1
* Reading from a dictionary doesn't need a lockJoshua Harlow2014-08-301-2/+1
* Use the same parent iterkeys method to be consistentJoshua Harlow2014-08-301-1/+2
* Mark disconnected before performing other activitiesJoshua Harlow2014-08-301-1/+1
* Keep the trailing slash only for sequenced createsJoshua Harlow2014-08-303-3/+8
* Use a constant to define what root isJoshua Harlow2014-08-301-4/+6
* Use a generator to ensure we get a unused sequence pathJoshua Harlow2014-08-302-7/+33
* Fix path postfix lost during normalizationJoshua Harlow2014-08-262-4/+16
* Remove useless helper functionJoshua Harlow2014-08-241-5/+1
* Move the partial client below the fake client codeJoshua Harlow2014-08-231-107/+109
* Add a counter multi-client validation testJoshua Harlow2014-08-221-0/+23
* Use retries and include the common client helpersJoshua Harlow2014-08-221-4/+31
* Correctly set data when a version is passedJoshua Harlow2014-08-221-0/+3
* Move _generate_async from client to utilsJoshua Harlow2014-08-222-23/+26
* Bump for 0.1 and include storage attached clientsJoshua Harlow2014-08-222-0/+18
* More work to make multi-clients work correctlyJoshua Harlow2014-08-184-73/+123
* Start to make multi-client storage possibleJoshua Harlow2014-08-173-5/+61