[8.x] Simulate exit code for closure scheduled tasks

[jszobody]

Currently, the onSuccess and onFailure hooks do not work for closure scheduled tasks. This is because these hooks rely on the exitCode, which is only tracked for external process calls.

That means in this case...

$schedule->call(function() {
    info('hello world');
})->onFailure(function() {
    info('task failed');
})->everyMinute();

... you will always find that the "task failed".

This PR aims to simulate a 0 (success) or 1 (general failure) exit code for closure tasks. An exit code of 1 is set if an exception is thrown OR false is returned from the closure. All other results are considered successful.

This allows the onSuccess and onFailure hooks (along with the related emailOutputOnFailure type methods) to work with closure tasks.

[mdevo]

Does this fixes applied on LTS also?

[driesvints]

@mdevo no this is a breaking change.

[ariaieboy]

in what version of laravel 8 this bug will get fix ?

[GrahamCampbell]

in what version of laravel 8 this bug will get fix ?

Every version of Laravel 8.

[GrahamCampbell]

GitHub tells you:

[ariaieboy]

i test this on laravel version 8.3.0 and 8.5.0 and its still not work
this is my scheduled job
$schedule->call(function (){ var_dump('test'); })->onFailure(function (){ var_dump('job failed'); })->onSuccess(function (){ var_dump('job succesd'); });
and its the output of my job
Running scheduled command: Closure string(4) "test" string(10) "job failed"

i tested it with closure and __invoke method in a class and with $schedule->job()
and in all this situations it always run the onFailure function;

[jszobody]

Confirmed. My mistake. Currently the callbacks are running before the exitCode is being set, looks like.

Line 90 here needs to be moved up, probably up to right after the $response is obtained inside the try block.

framework/src/Illuminate/Console/Scheduling/CallbackEvent.php

Lines 87 to 90 in 5a748ee

	parent::callAfterCallbacks($container);
	}
	$this->exitCode = $response === false ? 1 : 0;

@GrahamCampbell I can submit a new PR to fix this, will try to get to it soon.