You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@commons.apache.org by se...@apache.org on 2012/04/20 23:08:42 UTC
svn commit: r1328501 -
/commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
Author: sebb
Date: Fri Apr 20 21:08:41 2012
New Revision: 1328501
URL: http://svn.apache.org/viewvc?rev=1328501&view=rev
Log:
Document why we are not using the simpler Hex constants
Modified:
commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
Modified: commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
URL: http://svn.apache.org/viewvc/commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java?rev=1328501&r1=1328500&r2=1328501&view=diff
==============================================================================
--- commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java (original)
+++ commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java Fri Apr 20 21:08:41 2012
@@ -40,12 +40,17 @@ public class Precision {
* In IEEE 754 arithmetic, this is 2<sup>-53</sup>: {@value}.
*/
public static final double EPSILON = Double.longBitsToDouble((EXPONENT_OFFSET - 53l) << 52);
+ //This was previously expressed as = 0x1.0p-53;
+ // However, OpenJDK (Sparc Solaris) cannot handle such small constants: MATH-721
+
/**
* Safe minimum, such that {@code 1 / SAFE_MIN} does not overflow.
* In IEEE 754 arithmetic, this is also the smallest normalized
* number 2<sup>-1022</sup>: {@value}.
*/
public static final double SAFE_MIN = Double.longBitsToDouble((EXPONENT_OFFSET - 1022l) << 52);
+ // This was previously expressed as = 0x1.0p-1022;
+ // However, OpenJDK (Sparc Solaris) cannot handle such small constants: MATH-721
/** Offset to order signed double numbers lexicographically. */
private static final long SGN_MASK = 0x8000000000000000L;
Re: svn commit: r1328501 - /commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
Posted by sebb AT ASF <se...@apache.org>.
On 20 April 2012 22:08, <se...@apache.org> wrote:
> Author: sebb
> Date: Fri Apr 20 21:08:41 2012
> New Revision: 1328501
>
> URL: http://svn.apache.org/viewvc?rev=1328501&view=rev
> Log:
> Document why we are not using the simpler Hex constants
Although the reasoning was included in the commit message, anyone
reading the code would not see it, and might wonder why we did not use
the constants.
[Commit log messages are effectively ephemeral; they are really only
of use in understanding the commit email message.]
> Modified:
> commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
>
> Modified: commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java
> URL: http://svn.apache.org/viewvc/commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java?rev=1328501&r1=1328500&r2=1328501&view=diff
> ==============================================================================
> --- commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java (original)
> +++ commons/proper/math/trunk/src/main/java/org/apache/commons/math3/util/Precision.java Fri Apr 20 21:08:41 2012
> @@ -40,12 +40,17 @@ public class Precision {
> * In IEEE 754 arithmetic, this is 2<sup>-53</sup>: {@value}.
> */
> public static final double EPSILON = Double.longBitsToDouble((EXPONENT_OFFSET - 53l) << 52);
> + //This was previously expressed as = 0x1.0p-53;
> + // However, OpenJDK (Sparc Solaris) cannot handle such small constants: MATH-721
> +
> /**
> * Safe minimum, such that {@code 1 / SAFE_MIN} does not overflow.
> * In IEEE 754 arithmetic, this is also the smallest normalized
> * number 2<sup>-1022</sup>: {@value}.
> */
> public static final double SAFE_MIN = Double.longBitsToDouble((EXPONENT_OFFSET - 1022l) << 52);
> + // This was previously expressed as = 0x1.0p-1022;
> + // However, OpenJDK (Sparc Solaris) cannot handle such small constants: MATH-721
>
> /** Offset to order signed double numbers lexicographically. */
> private static final long SGN_MASK = 0x8000000000000000L;
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org