Fix json_decode problem

[theDisco]

As discussed in #1003 here is the PR that should at least make elastica to work properly when bigintConversion setting is set to true.

[ruflin]

This test would also pass without the setJonBigintVersion(true), right?

[theDisco]

Yes, this test would also pass without L215, but L215 makes the difference :)

[ruflin]

I'm fine with the test. But I assume this test would also have passed without the code changes you made? It is more for me to understand if we test if the bug was actually fixed or if we test that we can use the function without any side affects?

[theDisco]

No, it wouldn't since the return value of https://github.com/ruflin/Elastica/pull/1005/files#diff-b364251d02e70310f3c6e443372397d3L208 would be \stdClass and not an array and there is not conversion or casting happening till the end of method.

[ruflin]

@theDisco Thanks. Can you also update the CHANGELOG.md?

[theDisco]

@ruflin Sure, done.

[ruflin]

Sorry to bring this up again. But isn't the affect here, that if BIG_INT is enabled or not could be that the response object looks different?

[theDisco]

BIG_INT_* setting converts integers bigger than PHP_INT_MAX to a string, so yes, the response will be different (well some data types will be different), but it is the only way to deal with this without overflowing the integer.

[ruflin]

@theDisco Thanks for the contribution. We should definitively discuss if we should use JSON::parse($response, true) for the non JSON_BIG_INT conversion case. But I'm not sure if that will have any other side affects.

[theDisco]

@ruflin if you look at the implementation of JSON::parse() you actually do not have to pass true as a second parameter:

        // extract arguments
        $args = func_get_args();
        // default to decoding into an assoc array
        if (sizeof($args) === 1) {
            $args[] = true;
        }
        // run decode
        $array = call_user_func_array('json_decode', $args);

And no, passing true would not have any side effects.

[ruflin]

@theDisco Ugh :-( Surprising that nobody of us spotted that in the past. But that is great news.