Hacktoberfest 2026：维护者为十月标记出来的 issue，仍然开放、适合新手。 浏览 Hacktoberfest issue

Fragile window abandonment during concatMap operations; silent data loss

未关闭
  • 哪里不对？
#7,956 2 条评论 1 个 reaction 已指派 0 人 在 GitHub 查看

维护者通常 1 天内回复

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
38/100
Issue 类型
缺陷
描述清晰度
基本清楚
活跃度
停滞
技术栈
java
领域
backend-api-design

调研方向

首先使用 Flowable.window() 配合 concatMapSingle() 和 concatMapMaybe() 运行两个复现，并比较 blockingSubscribe() 的不同变体。跟踪窗口放弃与下游订阅之间的交互；当窗口不会被静默丢弃，或放弃会产生文档中说明的错误时，即表示完成。

由索引模型根据 Issue 内容生成。

描述

3.x 4.x Investigating

Combining the Flowable.window() family of operators with any of the Flowable.concatMap*() operators appears to cause window abandonment bugs which are sensitive to how the outer Flowable is processed downstream. This despite the given concatMap lambdas returning an Rx Single/Maybe ready to be subscribed as soon as they are returned and which will signal the upstream window for subscription synchronously.

In the demonstrations below, I subscribe to identical .window().concatMap() constructions with very slightly different downstreams (.blockingSuscribe(System.out::println) vs. .doOnNext(System.out::println).blockingSubscribe()) and they emit drastically different results due to silent data-loss.

Example: Windowing a stream into lists, similar to the buffer() operator.

    public static void main(String[] args) {
        final var content = Flowable
                .rangeLong(0, 100)
                .window(10)
                .concatMapSingle(Flowable::toList);

        System.out.println("Example 1: Expected behavior for concetMapSingle(Flowable::toList)");
        content
                .doOnNext(System.out::println)
                .blockingSubscribe();

        System.out.println("Example 2: Window abandonment & data-loss");
        content
                .blockingSubscribe(System.out::println);
    }

Output:

Example 1: Expected behavior for concetMapSingle(Flowable::toList)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
[10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
[20, 21, 22, 23, 24, 25, 26, 27, 28, 29]
[30, 31, 32, 33, 34, 35, 36, 37, 38, 39]
[40, 41, 42, 43, 44, 45, 46, 47, 48, 49]
[50, 51, 52, 53, 54, 55, 56, 57, 58, 59]
[60, 61, 62, 63, 64, 65, 66, 67, 68, 69]
[70, 71, 72, 73, 74, 75, 76, 77, 78, 79]
[80, 81, 82, 83, 84, 85, 86, 87, 88, 89]
[90, 91, 92, 93, 94, 95, 96, 97, 98, 99]

Example 2: Window abandonment & data-loss
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
[10]
[20]
[30]
[40]
[50]
[60]
[70]
[80]
[90]

Example: Windowing a stream into a simple reduction operation.

    public static void main(String[] args) {
        final var windowSums = Flowable
                .rangeLong(0, 100)
                .window(10)
                .concatMapMaybe(window -> window.reduce(Long::sum));

        System.out.println("Example 3: Expected behavior for concetMapMaybe(window -> window.reduce(Long::sum))");
        windowSums
                .doOnNext(System.out::println)
                .blockingSubscribe();

        System.out.println("Example 4: Silent data loss");
        windowSums
                .blockingSubscribe(System.out::println);
    }

Output:

Example 3: Expected behavior for concetMapMaybe(window -> window.reduce(Long::sum))
45
145
245
345
445
545
645
745
845
945

Example 4: Silent data loss
45
10
20
30
40
50
60
70
80
90

The flatMap*() family of operators seem less prone to this behavior but don't guarantee to retain the upstream ordering.

Ideally, I'd expect window() to be compatible with concatMap(). But if such is impossible, I'd expect the window-abandonment detection described in https://github.com/ReactiveX/RxJava/wiki/What%27s-different-in-3.0#window-abandonment-in-window to at least signal an error.

主要语言
Java
星标
48.2k
派生
7.6k
平均合并
3 小时 9 分钟
30 天内合并 PR
25

环境准备

从这里开始

  1. 先读完整个 Issue，再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库，在一个分支上完成修改。
  4. 提交 Pull Request，并在描述里引用这个 Issue 编号。

ReactiveX/RxJava 的其他 Issue

查看 ReactiveX/RxJava 的全部 Issue

相似的 Issue

更多 Java Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。