Version 1.0.5

[benjchristensen]

Preparing to release 1.0.5 ...

[benjchristensen]

@akarnokd @zsxwing What outstanding pull requests or issues do you want to make sure get into 1.0.5?

[akarnokd]

If all my open PRs could make it, that would be great.
2015.01.29. 0:23 ezt írta ("Ben Christensen" [email protected]):

@akarnokd https://github.com/akarnokd @zsxwing
https://github.com/zsxwing What outstanding pull requests or issues do
you want to make sure get into 1.0.5?

—
Reply to this email directly or view it on GitHub
#2555 (comment).

[benjchristensen]

Do you have any concerns with anything that has been merged?

Here is the diff from last release: v1.0.4...1.x

[akarnokd]

No.

[zsxwing]

I hope we can have some conclusion for #2549 before releasing 1.0.5, because I think awaitTerminalEvent(long timeout, TimeUnit unit) has been broken in #2332: before that it's quiet if timeout, but now it will throw a RuntimeException.

[akarnokd]

before that it's quiet if timeout, but now it will throw a RuntimeException

I think the correct behavior is what's now. Would you consider a test to be passed if it timed out silently?

[zsxwing]

I think the correct behavior is what's now. Would you consider a test to be passed if it timed out silently?

I think the current behavior is more reasonable. I just wonder the policy about changing the behavior of a method in TestObserver/TestSubscriber/TestScheduler. For the changing of awaitTerminalEvent(long timeout, TimeUnit unit), should it be classified as a bug fix or a breaking change?

[akarnokd]

These are test related code. You don't run unit tests in production, but only in development where you have access to your own tests and can update them according to the new behavior.

[benjchristensen]

I've reviewed all merged pull requests and am ready to release 1.0.5 except for a comment I added to #2465 about the naming convention of the system property.

[benjchristensen]

Here are the changes currently queued for release:

• Pull 2332 Operator retry test fix attempt
• Pull 2244 OperatorTakeLast add check for isUnsubscribed to fast path
• Pull 2469 Remove the execute permission from source files
• Pull 2455 Fix for Observable.cache() creates indefinite number of Threads with Schedulers.io() #2191 - OperatorMulticast fails to unsubscribe from source
• Pull 2474 MergeTest.testConcurrency timeout to let other tests run
• Pull 2335 A set of stateless operators that don't need to be instantiated
• Pull 2447 Fail early if a null subscription is added to a CompositeSubscription.
• Pull 2475 SynchronousQueue.clone fix
• Pull 2477 Backpressure tests fix0121
• Pull 2476 Fixed off-by-one error and value-drop in the window operator.
• Pull 2478 RefCountAsync: adjusted time values as 1 ms is unreliable
• Pull 2238 Fix the bug that cache doesn't unsubscribe the source Observable when the source is terminated
• Pull 1840 Unsubscribe when thread is interrupted
• Pull 2471 Fixes NPEs reported in TrampolineScheduler NullPointerException #1702 by synchronizing queue.
• Pull 2482 Merge: fixed hangs & missed scalar emissions
• Pull 2547 Warnings cleanup
• Pull 2465 ScheduledExecutorService: call purge periodically on JDK 6 to avoid
• Pull 2543 OperatorMerge handle request overflow
• Pull 2548 Subscriber.request should throw exception if negative request made
• Pull 2550 Subscriber.onStart requests should be additive (and check for overflow)
• Pull 2470 Experimental Subject state information methods & bounded ReplaySubject termination
• Pull 2553 RxRingBuffer with synchronization
• Pull 2540 Merge with max concurrency now supports backpressure.
• Pull 2565 Obstruction detection in tests.
• Pull 2563 Retry backpressure test: split error conditions into separate test lines.
• Pull 2572 Give more time to certain concurrency tests.
• Pull 2559 OnSubscribeFromIterable - add request overflow check
• Pull 2574 SizeEviction test needs to return false
• Pull 2561 Updating queue code from JCTools
• Pull 2566 CombineLatest: fixed concurrent requestUpTo yielding -1 requests
• Pull 2552 Publish: fixed incorrect subscriber requested accounting
• Pull 2583 Added perf tests for various container-like subscriptions
• Pull 2585 Experimental Operator: switchIfEmpty
• Pull 1955 OnBackpressureXXX: support for common drain manager & fix for former concurrency bugs

[akarnokd]

There are 3 bug fixes (2 trivial, 1 awaiting review) that should make it into 1.0.5. I've also created a PR to change the NewThreadWorker's system parameter names.

[benjchristensen]

Which of the 8 PRs should I look at?

[akarnokd]

Most urgent: https://github.com/ReactiveX/RxJava/pulls?q=is%3Aopen+is%3Apr+label%3ABug
Good to have soon: #2579
Can wait: the rest.

[benjchristensen]

Thanks, reviewing.

[benjchristensen]

Let's plan on releasing Wednesday if possible.

I apologize for letting 1.0.5 drag on so long ... I've been swamped in other areas recently. Thank you very much @akarnokd @zsxwing @davidmoten and others for all the work on 1.0.5.