Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #31855 > unrolled thread

Re: elementary loops

Back to article view | Back to comp.lang.java.programmer

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: elementary loops Marcel Mueller <news.5.maazl@spamgourmet.org> - 2015-06-26 09:17 +0200
    Re: elementary loops Andreas Leitgeb <avl@auth.logic.tuwien.ac.at> - 2015-06-26 09:26 +0000
      Re: elementary loops Marcel Mueller <news.5.maazl@spamgourmet.org> - 2015-06-26 13:39 +0200
        Re: elementary loops Eric Sosman <esosman@comcast-dot-net.invalid> - 2015-06-26 08:26 -0400
        Re: elementary loops Arne Vajhøj <arne@vajhoej.dk> - 2015-06-28 20:12 -0400
          Re: elementary loops Marcel Mueller <news.5.maazl@spamgourmet.org> - 2015-06-29 19:50 +0200
            Re: elementary loops Arne Vajhøj <arne@vajhoej.dk> - 2015-06-29 13:58 -0400

#31855 — Re: elementary loops

On 22.06.15 18.15, Stefan Ram wrote:
> public class Main
> { public static void main( final java.lang.String[] args )
>    { short i = -1;
>      int   j = -1;
>      for( int k = 0; k < 65; ++k ){ i >>>= 1; j <<= 32; }

The result of j<<=32 is a no-op and i>>>=1 is unsigned and should become 
zero.


Marcel

[toc] | [next] | [standalone]

#31856

Marcel Mueller <news.5.maazl@spamgourmet.org> wrote:
> On 22.06.15 18.15, Stefan Ram wrote:
>> public class Main
>> { public static void main( final java.lang.String[] args )
>>    { short i = -1;
>>      int   j = -1;
>>      for( int k = 0; k < 65; ++k ){ i >>>= 1; j <<= 32; }
> The result of j<<=32 is a no-op and i>>>=1 is unsigned and should become 
> zero.

The short i is promoted to int, then the MSB ("sign"-bit) zeroed,
then converted back to short, where it's still -1

[toc] | [prev] | [next] | [standalone]

#31857

On 26.06.15 11.26, Andreas Leitgeb wrote:
>>> { public static void main( final java.lang.String[] args )
>>>     { short i = -1;
>>>       int   j = -1;
>>>       for( int k = 0; k < 65; ++k ){ i >>>= 1; j <<= 32; }
>
> The short i is promoted to int, then the MSB ("sign"-bit) zeroed,
> then converted back to short, where it's still -1

In this case the >>> operator is meaningless for short int and there is 
no option to perform this operation except for emulating the behavior by 
other bit functions or types. The same applies to byte as well.
Of course, this is only another side effect of the missing unsigned types.
Is this guaranteed by the language or just implementation defined?


Marcel

[toc] | [prev] | [next] | [standalone]

#31858

On 6/26/2015 7:39 AM, Marcel Mueller wrote:
> On 26.06.15 11.26, Andreas Leitgeb wrote:
>>>> { public static void main( final java.lang.String[] args )
>>>>     { short i = -1;
>>>>       int   j = -1;
>>>>       for( int k = 0; k < 65; ++k ){ i >>>= 1; j <<= 32; }
>>
>> The short i is promoted to int, then the MSB ("sign"-bit) zeroed,
>> then converted back to short, where it's still -1
>
> In this case the >>> operator is meaningless for short int and there is
> no option to perform this operation except for emulating the behavior by
> other bit functions or types. The same applies to byte as well.
> Of course, this is only another side effect of the missing unsigned types.
> Is this guaranteed by the language or just implementation defined?

     The former.  The relevant bits of the Java Language Specification
<http://docs.oracle.com/javase/specs/jls/se8/html/index.html> are

     - 15.26.2 "Compound Assignment Operators"

     - 15.19 "Shift Operators"

     - 5.6.1 "Unary Numeric Promotion"

     - 5.1.2 "Widening Primitive Conversion"

     - 15.16 "Cast Expressions" (implied by 15.26.2)

     - 5.1.3 "Narrowing Primitive Conversion"

-- 
esosman@comcast-dot-net.invalid
"Don't be afraid of work. Make work afraid of you." -- TLM

[toc] | [prev] | [next] | [standalone]

#31867

On 6/26/2015 7:39 AM, Marcel Mueller wrote:
> On 26.06.15 11.26, Andreas Leitgeb wrote:
>>>> { public static void main( final java.lang.String[] args )
>>>>     { short i = -1;
>>>>       int   j = -1;
>>>>       for( int k = 0; k < 65; ++k ){ i >>>= 1; j <<= 32; }
>>
>> The short i is promoted to int, then the MSB ("sign"-bit) zeroed,
>> then converted back to short, where it's still -1
>
> In this case the >>> operator is meaningless for short int and there is
> no option to perform this operation except for emulating the behavior by
> other bit functions or types. The same applies to byte as well.
> Of course, this is only another side effect of the missing unsigned types.
> Is this guaranteed by the language or just implementation defined?

There is (luckily) very little implementation specific
and undefined in the Java language.

Arne

[toc] | [prev] | [next] | [standalone]

#31875

On 29.06.15 02.12, Arne Vajhøj wrote:
> There is (luckily) very little implementation specific
> and undefined in the Java language.

Oh, race conditions work pretty well in Java. Like in many other 
languages too, of course.


Marcel

[toc] | [prev] | [next] | [standalone]

#31876

On 6/29/2015 1:50 PM, Marcel Mueller wrote:
> On 29.06.15 02.12, Arne Vajhøj wrote:
>> There is (luckily) very little implementation specific
>> and undefined in the Java language.
>
> Oh, race conditions work pretty well in Java. Like in many other
> languages too, of course.

That is true.

But still way less of those things than f.ex. C.

Arne

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.java.programmer

csiph-web