Please review changes for JDK-8014519: scriptpad sample does not work with nashorn

[A. Sundararajan]

Hi Andreas,

I pushed your scriptpad patch to jdk8-tl. Thanks for your contribution!

-Sundar

On Tuesday 14 May 2013 10:16 PM, Attila Szegedi wrote:
> Thanks; +1 from me for committing this.
>
> On May 14, 2013, at 6:42 PM, Andreas Rieber <rieberandreas at gmail.com> wrote:
>
>> Answers inline...
>>
>> On 14.05.13 18:25, Attila Szegedi wrote:
>>> Hey Adreas,
>>>
>>> few things I'm curious about:
>>>
>>> -    if (! (callback instanceof Function)) {
>>> +    if (! (callback instanceof Function) && typeof callback !== "function") {
>>>
>>> When would you have a callback that is instanceof Function, but its typeof returns something other than "function"?
>> Patch is already some days old but as i remember it was Rhino/Nashorn differences and the case when someone would try:
>>
>>     setTimeout( "print('something')", 2000);
>>
>>> function mbean(objName, async) {
>>> +    var index;
>>>       objName = objectName(objName);
>>> -    var info = mbeanInfo(objName);
>>> +    var info = mbeanInfo(objName);
>>>       var attrs = info.attributes;
>>>       var attrMap = new Object;
>>> -    for (var index in attrs) {
>>> +    for (index in attrs) {
>>>           attrMap[attrs[index].name] = attrs[index];
>>>       }
>>>       var opers = info.operations;
>>>       var operMap = new Object;
>>> -    for (var index in opers) {
>>> +    for (index in opers) {
>>>           operMap[opers[index].name] = opers[index];
>>>       }
>>>
>>> What do we gain from extracting "var index" from loop declarations into a function-level variable?
>> The 'var index' has function scope, so a second 'var index' is wrong. There is no block scope in javascript ;-)
>>
>>>   // allocate an integer array of "big enough" size!
>>>   var a = java.lang.reflect.Array.newInstance(
>>> -            java.lang.Integer.TYPE, 1024*1024);
>>> -
>>> -// loop forever!
>>> -while (true);
>>> +    java.lang.Integer.TYPE, input * 1024 * 1024);
>>>
>>> I know this doesn't pertain to the reason for this being in the patch, just wanted to note that in Nashorn you can use Java.type() to obtain the type for int[], and you can subsequently use that as a constructor to instantiate a new Java int array:
>>>
>>> var a = new (Java.type("int[]"))(input * 1024 * 1024)
>>>
>>> You don't have to go through java.lang.reflect.Array.
>> True, i didn't change that to keep Nashorn and Rhino working. Now it could be changed the Nashorn way.
>>
>> cheers
>> Andreas
>>
>>> Regards,
>>>    Attila.
>>>
>>> On May 14, 2013, at 5:13 PM, "A. Sundararajan" <sundararajan.athijegannathan at oracle.com> wrote:
>>>
>>>> Please review http://cr.openjdk.java.net/~sundar/8014519/
>>>>
>>>> This is a contribution by Andreas Rieber
>>>>
>>>> Thanks
>>>> -Sundar
>>