RFR: 8321124: java/util/stream/GatherersTest.java times out [v4]

[Alan Bateman]

On Mon, 4 Dec 2023 14:25:05 GMT, Viktor Klang <vklang at openjdk.org> wrote:

>> Splits GatherersTest out to distinct files per built-in gatherer
>> 
>> Fixes a test ordering issue for mapConcurrent which placed GatherersTest on the ProblemList.
>> 
>> Adding increased maxOutputSize for Gatherer-related tests to improve debuggability.
>> 
>> Lowering the composition threshold of GathererTest.testMassivelyComposedGatherers to 256 to avoid SOE on low-specc:ed machines.
>
> Viktor Klang has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Removing extraneous built-in gatherer test output

test/jdk/java/util/stream/GatherersMapConcurrentTest.java line 106:

> 104:             );
> 105:         assertEquals("expected", exception.getMessage());
> 106:         assertNull(exception.getCause());

I assume it would be possible to add a comrade to this test for the wrapping case too, e.g. if the mapper throws an Error.

test/jdk/java/util/stream/GatherersMapConcurrentTest.java line 245:

> 243:     public void behavesAsExpected(Config config) {
> 244:         for (var concurrency : List.of(1, 2, 3, 10, 1000)) {
> 245:             final var expectedResult = config.stream()

I assume the loop is here because you can't have both a MethodSource and ValueSource. I think I'd probably create a new method source for this, e.g.


    static Stream<Arguments> configutatonAndConcurrency() {
        return IntStream.of(1, 2, 3, 10, 1000)
                .mapToObj(i -> i)
                .flatMap(i -> configurations().map(c -> new Arguments(c, i)));
    }


The advantage is that you'll have the parameter in the .jtr in the event of a failure.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/16928#discussion_r1415671564
PR Review Comment: https://git.openjdk.org/jdk/pull/16928#discussion_r1415665826