GNU Scientific Library - Bugs: bug #23247, gsl_complex_inverse for real gives...
You are not allowed to post comments on this tracker with your current authentication level.
bug #23247: gsl_complex_inverse for real gives -0 for imaginary part
Submitter: | -Deleted Account- <bjg> | ||
Submitted: | Tue 13 May 2008 05:48:06 PM UTC | ||
Category: | Accuracy problem | Severity: | 2 - Minor |
Operating System: | Status: | Invalid | |
Assigned to: | None | Open/Closed: | Closed |
Release: | 1.11 |
Tue 18 Nov 2008 07:47:32 PM UTC, comment #2: |
-Deleted Account- <bjg> |
Tue 13 May 2008 05:49:16 PM UTC, comment #1: Need to check what Kahan IEEE complex branch paper says on 1/z with +-0 for imag. |
-Deleted Account- <bjg> |
Tue 13 May 2008 05:48:06 PM UTC, original submission:
From: Chris <chris.colemansmith@gmail.com>
|
-Deleted Account- <bjg> |
Attached Files
Depends on the following items: None found
Items that depend on this one: None found
CC list is empty
Follow 3 latest changes.
Date | Changed by | Updated Field | Previous Value | => | Replaced by |
---|---|---|---|---|---|
2008-11-18 | bjg | Status | None | Invalid | |
Open/Closed | Open | Closed | |||
2008-05-13 | bjg | Attached File | - | Added negz.c, #15660 |
Powered by Savane 3.13-d3ae.
Corresponding source code
Given that the inverse of x+iy is (x-iy)/|z|^2 I believe that returning the reciprocal of 1/(x+i0) as (1/x)-i0 is correct (i.e. when looked at as a limiting value).