Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!news.chainon-marquant.org!nntpfeed.proxad.net!proxad.net!feeder2-2.proxad.net!newsfeed.arcor.de!newsspool3.arcor-online.net!news.arcor.de.POSTED!not-for-mail
Content-Type: text/plain; charset="UTF-8"
Message-ID: <1932615.aqSPkdTlGX@PointedEars.de>
From: Thomas 'PointedEars' Lahn <PointedEars@web.de>
Reply-To: Thomas 'PointedEars' Lahn <cljs@PointedEars.de>
Organization: PointedEars Software (PES)
Date: Thu, 29 Dec 2011 12:18 +0100
User-Agent: KNode/4.4.11
Content-Transfer-Encoding: 8Bit
Subject: Re: Simulate goto
Newsgroups: comp.lang.javascript
References: <ef9039a7-0cf7-4887-a98e-81c49968555b@p42g2000vbt.googlegroups.com> <4ef9a823$0$6842$e4fe514c@news2.news.xs4all.nl> <1931071.aqSPkdTlGX@PointedEars.de> <OnnMPUAzo0+OFwwl@J.A830F0FF37FB96852AD08924D9443D28E23ED5CD> <1367331.WGnKmheAe9@PointedEars.de> <9m1iteFodiU1@mid.uni-berlin.de>
Followup-To: comp.lang.javascript
MIME-Version: 1.0
Lines: 102
NNTP-Posting-Date: 29 Dec 2011 12:18:00 CET
NNTP-Posting-Host: e0268169.newsspool1.arcor-online.net
X-Trace: DXC=ZGaKCi7JJd7T2Rfi6<W9f5ic==]BZ:af>4Fo<]lROoR1<`=YMgDjhg2T`38E:`Ooj8DZm8W4\YJN<I3o8e0@o0D;4fl4=K?<S<6XFa4^l[[_6>
X-Complaints-To: usenet-abuse@arcor.de
Xref: x330-a1.tempe.blueboxinc.net comp.lang.javascript:9759

Jake Jarvis wrote:

> On 28.12.2011 19:10, Thomas 'PointedEars' Lahn wrote:
>> AIUI, it should be possible to use an enclosing /Statement/ (like a
>> /Block/ statement) and a `break' statement here (section 12.8):
>>
>>    function testGoto()
>>    {
>>      var isFirst = true;
>>      if (true)
>>      {
>> skipPoint:
>>        window.alert("Using goto\n");
>>
>>        if (isFirst)
>>        {
>>          isFirst = false;
>>          break skipPoint;
>>          window.alert("This part is skipped");
>>        }
>>      }
>>    }
>>
>> However, Google V8 3.5.10.9+ in Chromium 15.0.874.121 (Developer Build
>> 109964 Linux) considers that a syntax error as well (same error message:
>> "SyntaxError: Undefined label 'skipPoint'").  Is this a bug?
> 
> No.
> Why should it work?
> I think there is slight confusion as to what it means for a statement to
> be 'enclosing' another.

No, but there was a misconception on my part as to the definition of a label 
set:

| 12 Statements
| 
| Semantics
| 
| A /Statement/ can be part of a /LabelledStatement/, which itself can be
| part of a /LabelledStatement/, and so on. The labels introduced this way
| are collectively referred to as the "current label set" when describing
| the semantics of individual statements. A /LabelledStatement/ has no
| semantic meaning other than the introduction of a label to a /label set/.
| The label set of an /IterationStatement/ or a /SwitchStatement/ initially
| contains the single element empty. The label set of any other statement is
| initially empty.
|
| […]
| 12.12 Labelled Statements 
| 
| The production /Identifier/ : /Statement/ is evaluated by adding
| /Identifier/ to the label set of /Statement/ and then evaluating
| /Statement/. If the /LabelledStatement/ itself has a non-empty label set,
| these labels are also added to the label set of /Statement/ before
| evaluating it. If the result of evaluating Statement is (break, V, L)
| where L is equal to /Identifier/, the production results in (normal, V,
| empty).
| 
| Prior to the evaluation of a /LabelledStatement/, the contained
| /Statement/ is regarded as possessing an empty label set, unless it is an
| /IterationStatement/ or a /SwitchStatement/, in which case it is regarded
| as possessing a label set consisting of the single element, empty.

That is, the label set of a statement consists of the labels used for *that* 
statement, _not_ the labels used *within* it.  Therefore, the following is 
syntactically correct and compiles in the named implementation without 
error:

  function testGoto()
  {
    var counter = 0;
    
    skipPoint:
    {
      window.alert(counter);

      if (counter === 0)
      {
        ++counter;
        break skipPoint;
        window.alert("This part is skipped");
      }
    }
  }

However, this does _not_ "simulate goto", i. e. the counter is only shown 
once.  That is – unless I am overlooking something – intended by the 
Specification: The `break' statement with /Identifier/ merely ends the 
evaluation of the so labelled statement prematurely:

| A /BreakStatement/ with an /Identifier/ is evaluated as follows:
| 
| 1. Return (break, empty, /Identifier/).

See section 12.12 quoted above on how that result affects the evaluation of 
the enclosing statement labelled with /Identifier/.


PointedEars
-- 
When all you know is jQuery, every problem looks $(olvable).