Question

我有以下重试逻辑来重试操作。它适用于单个请求。对于多个正在进行的请求，我想在重试之前等待现有的重试逻辑完成。

handleError(errors: Observable<any>) {

    const retryCountStart: number = 1;

    // wait if there is any existing operation retrying
    // once it is complete, continue here

    return errors
        .mergeScan<any, any>(
        (retryCount: any, err: any) => {

            if (retryCount <= 5) {
                return Observable.of(retryCount + 1);
            } 

        },retryCountStart)
        .delay(1000);
}

如何在上述方法中满足某些条件之前添加延迟？

Answer 1

您可以将async / await用于Promise解决方案：

async handleError(errors: Observable<any>) {

    const retryCountStart: number = 1;

    // wait if there is any existing operation retrying
    // ----------------------------------------------------------
    await new Promise(resolve => {
        // declare some global variable to check in while loop
        while(this.retrying){
            setTimeout(()=> {
                // Just adding some delay 
                // (you can remove this setTimeout block if you want)
            },50);
        }

        // when while-loop breaks, resolve the promise to continue
        resolve();
    });
    // ----------------------------------------------------------

    // once it is complete, continue here

    return errors
        .mergeScan<any, any>(
        (retryCount: any, err: any) => {

            if (retryCount <= 5) {
                return Observable.of(retryCount + 1);
            } 

        },retryCountStart)
        .delay(1000);
}

Answer 2

据我了解，您只想在前一个流完成（即将流添加到队列）之后才开始下一个流

import { Observable, of, BehaviorSubject, from } from 'rxjs';
import { tap, finalize, filter, take, switchMap, delay } from 'rxjs/operators';

class StreamQueue {
  lastStreamCompleted$: Observable<boolean> = new BehaviorSubject(true);

  private runAfter<T>(lastStreamCompleted$: Observable<boolean>, stream$: Observable<T>): [Observable<boolean>, Observable<T>] {
    const newLastStreamCompleted$ = new BehaviorSubject(false);
    const newStream$ = lastStreamCompleted$
      .pipe(
        filter(lastStreamCompleted => lastStreamCompleted),
        take(1),
        switchMap(() => stream$),
        finalize(() => newLastStreamCompleted$.next(true)),
    );
    return [newLastStreamCompleted$, newStream$];
  }

  add(stream$: Observable<any>) {
    const [newLastStreamCompleted$, newStream$] = this.runAfter(this.lastStreamCompleted$, stream$);
    this.lastStreamCompleted$ = newLastStreamCompleted$;
    return newStream$;
  }
}

const streamQueue = new StreamQueue();

streamQueue.add(from([1, 2]).pipe(delay(100))).subscribe(console.log);
setTimeout(()=>streamQueue.add(from([21, 22]).pipe(delay(100))).subscribe(console.log), 100);
streamQueue.add(from([11, 12]).pipe(delay(100))).subscribe(console.log);

// Output:
// 1
// 2
// 11
// 12
// 21
// 22

Rxjs可观察到等待，直到满足某些条件

2 个答案: