[Octave-bug-tracker] [bug #61787] wrong output of "isreal" on array of uint8

[Hartmut]

URL:
  <https://savannah.gnu.org/bugs/?61787>

                 Summary: wrong output of "isreal" on array of uint8
                 Project: GNU Octave
            Submitted by: hardy
            Submitted on: Mon 10 Jan 2022 05:17:23 PM UTC
                Category: Octave Function
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Regression
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
                 Release: 7.0.90
         Discussion Lock: Any
        Operating System: Any

    _______________________________________________________

Details:

Here is a regression (I think) that I observered with the current release
candidate 7.0.90 under Win10, but I assume it is there under all OS'es. To
demonstrate the principle, run the following code:


isreal(uint8(5):uint8(9))


* This gives 0 under Octave 7.0.90
* This gives 1 under Octave 6.4.0
* This gives 1 under Matlab R2019b

I think the current result of Octave 7.0.90 is wrong. Maybe the underlying
regression in Octave 7 is much wider than just the "isreal" function, I cannot
judge this.

This error does not happen if you just do "isreal(5:9)" with double input
arrays.

I found this, because a unit test in the image package was failing.




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?61787>

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