| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Forgot to remove this in https://github.com/dpkp/kafka-python/pull/1925
/ ca2d76304bfe3900f995e6f0e4377b2ef654997e
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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* Old partitioners used by `SimpleProducer`; these are superceded by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`.
|
|
|
|
|
| |
Snappy URL was outdated. Similarly, many of these sites now support
https.
|
| |
|
|
|
| |
https://github.com/dpkp/kafka-python/pull/1304/files#r155842880
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
* Use weakref when registering a producer.close atexit to fix normal gc
* Test that del(producer) terminates async thread
|
| |
|
| |
|
| |
|
|
|
|
| |
encoding spec
|
|
|
|
| |
- previously this would simply drop one of the payloads
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
It will still die, just as before, but it now includes a *helpful* error message
|
| |
|
|
|
|
| |
in memory logging. Address code review concerns
|
|
|
|
|
|
|
|
| |
Bump version number to 0.9.1
Update readme to show supported Kafka/Python versions
Validate arguments in consumer.py, add initial consumer unit test
Make service kill() child processes when startup fails
Add tests for util.py, fix Python 2.6 specific bug.
|
| |
|
| |
|
|
|
|
|
|
| |
Conflicts:
kafka/util.py
|
|
|
|
| |
Set FetchRequest MaxBytes value to bufsize instead of fetchsize (=MinBytes)
|
| |
|
|
|
|
| |
Also move the exceptions to common instead of util
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Related to #42
Adds new ConsumerFetchSizeTooSmall exception that is thrown when
`_decode_message_set_iter` gets a BufferUnderflowError but has not yet
yielded a message
In this event, SimpleConsumer will increase the fetch size by 1.5 and
continue the fetching loop while _not_ increasing the offset (basically
just retries the request with a larger fetch size)
Once the consumer fetch size has been increased, it will remain
increased while SimpleConsumer fetches from that partition
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous commit optimized the commit thread such that the timer
started only when there were messages to be consumed. This commit
goes a step further and ensures the following:
* Only one timer thread is created
* The main app does not block on exit (waiting for timer thread to finish)
This is ensured by having a single thread blocking on an event and
keeps calling a function. We use events instead of time.sleep() so
as to prevent the python interpreter from running every 50ms checking
if the timer has expired (logic copied from threading.Timer)
|
|
|
|
|
|
|
| |
If there are no messages being consumed, the timer keeps
creating new threads at the specified intervals. This may
not be necessary. We can control this behaviour such that
the timer thread is started only when a message is consumed
|
|
|
|
| |
consumer.py and conn.py will be done later after pending merges
|
|
|
|
|
|
|
| |
The auto commit timer is one-shot. After the first commit, it does
not fire again. This ticket fixes the issue.
Also, in util.ReentrantTimer(), some duplicate code was cleaned up
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Also added a bunch of docstrings
|
|
|