summaryrefslogtreecommitdiff
path: root/kafka/producer/simple.py
Commit message (Collapse)AuthorAgeFilesLines
* Remove SimpleClient, Producer, Consumer, Unittest (#1196)Jeff Widman2019-10-111-54/+0
| | | | | | | | | | | | | | | | | | In the 2.0 release, we're removing: * `SimpleClient` * `SimpleConsumer` * `SimpleProducer` * Old partitioners used by `SimpleProducer`; these are superceded by the `DefaultPartitioner` These have been deprecated for several years in favor of `KafkaClient` / `KafkaConsumer` / `KafkaProducer`. Since 2.0 allows breaking changes, we are removing the deprecated classes. Additionally, since the only usage of `unittest` was in tests for these old Simple* clients, this also drops `unittest` from the library. All tests now run under `pytest`.
* Be explicit with tuples for %s formattingJeff Widman2018-11-181-1/+1
| | | | Fix #1633
* Vendor `six` consistentlyJeff Widman2018-10-221-2/+2
| | | | Use vendored `six`, and also `six.moves.range` rather than `xrange`
* Change SimpleProducer to use async_send (async is reserved in py37) (#1454)Dana Powers2018-03-231-1/+1
|
* use absolute imports everywhere (#1362)Kevin Tindall2018-02-061-1/+1
|
* Vendor six 1.10.0sixDana Powers2016-08-011-1/+1
|
* Remove unused importskafka-3318Dana Powers2016-04-051-1/+0
|
* Merge branch '0.9'Dana Powers2016-01-071-3/+0
|\ | | | | | | | | | | | | | | Conflicts: kafka/codec.py kafka/version.py test/test_producer.py test/test_producer_integration.py
| * Drop kafka_bytestringDana Powers2015-12-101-3/+0
| |
* | Handle new topic creation / LeaderNotAvailableError in initial ↵Dana Powers2015-12-161-1/+1
|/ | | | SimpleProducer.send_messages call
* Dont maintain all producer args / kwargs in subclass __init__ and docstrings ↵Dana Powers2015-06-081-48/+14
| | | | -- just refer to super class (Producer)
* Deprecate async producer batch_send kwarg -- use 'async' insteadDana Powers2015-06-061-2/+2
|
* Use separate module loggers instead of a single 'kafka' loggerDana Powers2015-06-061-1/+1
|
* Improve async producer code: logic and style fixesViktor Shlapakov2015-06-031-4/+8
| | | | | | | | | | - send_producer_request with fail_on_error=False to retry failed reqs only - using an internal dict with with namedtuple keys for retry counters - refresh metadata on refresh_error irrespective to retries options - removed infinite retries (retry_options.limit=None) as an over-feature - separate producer init args for retries options (limit,backoff,on_timeouts) - AsyncProducerQueueFull returns a list of failed messages - producer tests improved thanks to @rogaha and @toli
* Fix names for async retries opts, add timeout for putViktor Shlapakov2015-06-031-3/+8
|
* Improved retry logicViktor Shlapakov2015-06-031-6/+3
|
* Retries for async batchingViktor Shlapakov2015-06-031-3/+8
|
* Just encode to bytes if it isn't bytes.John Anderson2015-01-241-2/+4
|
* Alert if the topic is not sent as bytesJohn Anderson2015-01-231-0/+4
|
* Merge pull request #282 from wedaly/sphinx-api-docsDana Powers2015-01-231-14/+16
|\ | | | | Add Sphinx API docs
| * Add Sphinx API docsWill Daly2015-01-151-14/+16
| |
* | Randomize start by default for SimpleProducerAlex Couture-Beil2015-01-091-1/+1
|/ | | | | | | | random_start=False is dangerous. Any client that initializes a SimpleProducer and performs only a single publish batch will always go to partition 0. A common use-case for this is command line utilities, or web-servers which are unable to cache the SimpleProducer instance between calls.
* Separate consumers/producers/partitionersDana Powers2014-09-101-0/+73