Review request for JDK-8006984

[Andreas Rieber]

absolutely correct, and easy to oversee!

On 25.02.13 20:19, Attila Szegedi wrote:
> (Apologies if this came across overly nit-picking and pedantic; I just 
> wanted to make sure that I have a detailed trail of my reasoning out 
> in the open)
>
> On Feb 25, 2013, at 8:18 PM, Attila Szegedi <attila.szegedi at oracle.com 
> <mailto:attila.szegedi at oracle.com>> wrote:
>
>> If you look at the Program grammar production in Chapter 14, it says:
>>
>> /Program /*:*
>>
>> /SourceElements/_opt
>>
>> /SourceElements /*:*
>>
>> /SourceElement
>> SourceElements SourceElement/
>>
>> /SourceElement /*:*
>>
>> /Statement <http://es5.github.com/#x12>
>> FunctionDeclaration <http://es5.github.com/#x13>/
>> As you can see, SourceElement is either a Statement or a 
>> FunctionDeclaration -- they're considered two different things, and 
>> through their union SourceElement they can both occur as top-level 
>> elements in a program.
>>
>> Further, the grammar production FunctionBody in Chapter 13 is:
>>
>> /FunctionBody /*:*
>>
>> /SourceElements <http://es5.github.com/#x14>/_opt
>> meaning a function body also consists of SourceElements - either 
>> Statements or FunctionDeclarations. That are the only contexts in 
>> which FunctionDeclaration can occur, it occurs nowhere else in the 
>> grammar.
>>
>> Finally, Chapter 12 "Statements" is quite articulate in *not* listing 
>> FunctionDeclaration as part of the Statement production:
>>
>> /Statement /*:*
>>
>> /Block
>> VariableStatement
>> EmptyStatement
>> ExpressionStatement
>> IfStatement
>> IterationStatement
>> ContinueStatement
>> BreakStatement
>> ReturnStatement
>> WithStatement
>> LabelledStatement
>> SwitchStatement
>> ThrowStatement
>> TryStatement
>> DebuggerStatement/
>> There is also this non-normative "NOTE" section in Chapter 12 saying 
>> "Several widely used implementations of ECMAScript are known to 
>> support the use of FunctionDeclaration as a Statement. However there 
>> are significant and irreconcilable variations among the 
>> implementations in the semantics applied to such 
>> FunctionDeclarations. Because of these irreconcilable differences, 
>> the use of a FunctionDeclaration as a Statement results in code that 
>> is not reliably portable among implementations. …"
>>
>> The specification writers clearly frowns upon this practice, for 
>> quite obvious reasons - function declarations are instantiated upon 
>> entry to the program or their parent function, not when the 
>> declaration is executed in source code statement order, therefore 
>> programmers can have mistaken assumptions about their lexical scope, 
>> because such declarations always have the VariableEnvironment of the 
>> execution context at script/function entry point as their scope, and 
>> not the LexicalEnvironment at the point of their declaration (which 
>> is what function expressions get).
>>
>> We hope that by strict adherence to the specification we can steer 
>> people towards writing cleaner, more maintainable code in the future; 
>> one of goals of the ECMAScript 5.1 specification was eliminating or 
>> clarifying a lot of arcane quirks of the language; this purpose would 
>> be partially defeated if we just ignored it.
>>
>> HTH,
>>   Attila.
>>
>>
>> On Feb 25, 2013, at 8:03 PM, Andreas Rieber <rieberandreas at gmail.com 
>> <mailto:rieberandreas at gmail.com>> wrote:
>>
>>> Hi Attila,
>>>
>>> i can't read that out of ecma-262 spec but i guess you read that 
>>> many times more than i did. I try to follow "The Good Pars", which 
>>> would make all easier.
>>>
>>> Is there a way to access the issue tracking? I mean it would make my 
>>> live easier. At the moment i follow every commit and try to get in 
>>> but i don't see what is open, coming, etc.
>>>
>>> cheers
>>> Andreas
>>>
>>> PS: if i am not totally wrong - that code was from Sundar.
>>>
>>>
>>> On 25.02.13 18:46, Attila Szegedi wrote:
>>>> Excellent…
>>>>
>>>> On that note, I want to make you aware that we'll soon make Nashorn 
>>>> reject with a syntax error those function declarations that occur 
>>>> within "with" and "catch" blocks; we're shooting for 100% 
>>>> ECMAScript 5.1 compliance, and it disallows function declarations 
>>>> except directly on the top-level of script and in another 
>>>> function's body. In such cases, you'll need to use a function 
>>>> expression instead. So this:
>>>>
>>>> var guiPkgs = { JFrame: function() { print("created"); } };
>>>>
>>>> with (guiPkgs) {
>>>>       function main() { // <---- this is not legal in ECMAScript 5.1
>>>>         var frame;
>>>>
>>>> function createFrame() {
>>>> frame = new JFrame();
>>>>         }
>>>>
>>>> createFrame();
>>>>     }
>>>> }
>>>> main();
>>>>
>>>> will soon stop working. If you want the "main" function and 
>>>> whatever is in it to be affected by the with() block, you'll have 
>>>> to use a function expression instead like this:
>>>>
>>>> with (guiPkgs) {
>>>>       var main = function() { // <---- this expresses your actual 
>>>> intent, and is the only valid construct here in ES5.1
>>>>         var frame;
>>>>
>>>> function createFrame() { // <-- this is still okay, as it's defined 
>>>> directly on the function body level of main()
>>>> frame = new JFrame();
>>>>         }
>>>>
>>>> createFrame();
>>>>     }
>>>> }
>>>> main();
>>>>
>>>> Attila.
>>>>
>>>> On Feb 25, 2013, at 6:37 PM, Andreas Rieber 
>>>> <rieberandreas at gmail.com <mailto:rieberandreas at gmail.com>> wrote:
>>>>
>>>>> Hi Attila,
>>>>>
>>>>> i found that fix and retested. One line more and another one... 
>>>>> couldn't break it ;-)
>>>>>
>>>>> Andreas
>>>>>
>>>>> On 25.02.13 16:36, Attila Szegedi wrote:
>>>>>> Second review request, now with the test: 
>>>>>> http://cr.openjdk.java.net/~attila/8006984/webrev.01/ 
>>>>>> <http://cr.openjdk.java.net/%7Eattila/8006984/webrev.01/>
>>>>>>
>>>>>> On Feb 25, 2013, at 3:43 PM, Attila Szegedi 
>>>>>> <attila.szegedi at oracle.com <mailto:attila.szegedi at oracle.com>> wrote:
>>>>>>
>>>>>>> Please review JDK-8006984 at 
>>>>>>> http://cr.openjdk.java.net/~attila/8006984/webrev.00 
>>>>>>> <http://cr.openjdk.java.net/%7Eattila/8006984/webrev.00>
>>>>>>>
>>>>>>> Thanks,
>>>>>>>  Attila.
>>>>