[Octave-bug-tracker] [bug #62817] Regression in interpn, likely harmless

[Arun Giridhar]

URL:
  <https://savannah.gnu.org/bugs/?62817>

                 Summary: Regression in interpn, likely harmless
                 Project: GNU Octave
               Submitter: arungiridhar
               Submitted: Tue 26 Jul 2022 08:36:07 PM EDT
                Category: None
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: None
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
                 Release: dev
         Discussion Lock: Any
        Operating System: GNU/Linux


    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Tue 26 Jul 2022 08:36:07 PM EDT By: Arun Giridhar <arungiridhar>
Getting a new regression after a recent distro upgrade:

>> test interpn
***** test <*61907>
 yi = [0.5, 1.5]';
 xi = [2.5, 3.5];
 zi = [2.25, 4.75];
 v = rand (4, 3, 5) + 1i * rand (4, 3, 5);
 for method = {"nearest", "linear", "spline"}
   vi_complex = interpn (v, yi, xi, zi, method{1});
   vi_real = interpn (real (v), yi, xi, zi, method{1});
   vi_imag = interpn (imag (v), yi, xi, zi, method{1});
   assert (real (vi_complex), vi_real)
   assert (imag (vi_complex), vi_imag)
 endfor
!!!!! regression: https://octave.org/testfailure/?61907
ASSERT errors for:  assert (real (vi_complex),vi_real)

  Location  |  Observed  |  Expected  |  Reason
  (2,1,2)      0.46978      0.46978      Abs err 5.5511e-17 exceeds tol 0 by
6e-17
PASSES 18 out of 19 tests


It's probably nothing serious. Is it OK to increase the tolerance on that test
to eps instead of 0?








    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?62817>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/