Possible bug in no-nested-subscribe

[fchiumeo]

The no-nested-subscribe rule should only report subscribe within a subscribe callback or within any operator callback ?
Because it's currently reporting an error in a subscribe within a map callback
(in package rxjs-tslint-rules it's OK)

from(/* Observable */)
    .pipe(
        map((items: any[]) =>
        {
            form.get('control').valueChanges
                //  error  Nested subscribe calls are forbidden  rxjs/no-nested-subscribe
                .subscribe((billing: boolean) => 
                {
                    // code
                });
        }))
    .subscribe(() =>
    {
        //more code
    });

[cartant]

Nope. The rule is working as intended. Explicit subscribe calls within composed observables should be avoided.

Hmm. Maybe not. Looking at the tests, it's specifically about subscribe calls within subscribe, so despite the snippet being something that'd I would avoid, it's not something for which the rule is supposed to effect failures.

But, TBH, I'm more included to add more tests - for this behaviour - than I am to remove it.

[cartant]

This should be fixed in 3.1.5. I was mistaken and it is a bug, because usage like this is fine:

eslint-plugin-rxjs/tests/rules/no-nested-subscribe.ts

Lines 64 to 68 in 213794b

	// https://github.com/cartant/eslint-plugin-rxjs/issues/67
	import { Observable, of } from "rxjs";
	new Observable<number>(subscriber => {
	of(42).subscribe(subscriber);
	}).subscribe();

However, usage in callbacks passed to operators is probably still a problem. At some stage, I might extend the rule to effect failures for explicit subscribe calls in functions passed to operators.