bugGNU Octave - Bugs: bug #52699, ginput() broken on dev


bug #52699: ginput() broken on dev

Submitted by:  Rik <rik5>
Submitted on:  Tue 19 Dec 2017 09:28:06 PM UTC  
Category: Octave FunctionSeverity: 3 - Normal
Priority: 5 - NormalItem Group: Regression
Status: FixedAssigned to: None
Originator Name: Open/Closed: Closed
Release: devOperating System: GNU/Linux

Add a New Comment(Rich Markup)

You are not logged in

Please log in, so followups can be emailed to you.


(Jump to the original submission Jump to the original submission)

Tue 26 Dec 2017 06:37:13 PM UTC, comment #9:

I added tests for emitting an error if a function parameter is declared persistent here http://hg.savannah.gnu.org/hgweb/octave/rev/2dc926274bce. These are part of the fixed tests in the test/ directory.

I tried the equivalent with global (declaring the input or output of a function global) and there is no warning in Octave. Should we at least warn about that as Matlab does?

Rik <rik5>
Project Administrator
Mon 25 Dec 2017 07:47:19 PM UTC, comment #8:

Could we try your suggestion to not allow existing
variables to be made global?

If there is complaint from Matlab users other choices could
be considered, or??

After a moment's thought, I do not think that flag to allow
global of existing is a good idea at all.

Michael Godfrey <godfrey>
Project Member
Sun 24 Dec 2017 05:21:07 PM UTC, comment #7:

This raises a question that I have wondered about: Is there
a case for a "Matlab compatibility" mode? This could be
set by some variable or option.

However, it would seem best to disallow making existing
variables or function parameters global. Use of global is
in any case prone to problems. (Even though I have used
it myself, I know that I should not.)

Michael Godfrey <godfrey>
Project Member
Sun 24 Dec 2017 05:00:27 PM UTC, comment #6:

Matlab allows function parameters and existing variables to be made global, but there is a warning that the local value may change to match the global value and that this behavior may go away in some future release. Who knows when that might be though. I mean, it has been documented for many many years that isstr will go away and that you should use ischar instead, but that hasn't happened yet.

I would prefer that we disallow making existing variables (including function parameters) global the same as we do for persistent, but we would probably get complaints about how that behavior is not compatible with Matlab.

John W. Eaton <jwe>
Project Administrator
Sun 24 Dec 2017 01:03:24 AM UTC, comment #5:

Solution works for me. Should there be a note in the manual about how persistent variables can't be input/output parameters for functions?

Does the restriction also apply to global variables?

Rik <rik5>
Project Administrator
Sat 23 Dec 2017 07:35:16 PM UTC, comment #4:

I checked in the following changeset:


Next, I'll see if I can make Octave throw an error if persistent is applied to function arguments (in or out). I think this new behavior is compatible with Matlab. Also, I don't really understand how it ever worked before. Hmm.

John W. Eaton <jwe>
Project Administrator
Sat 23 Dec 2017 12:23:55 AM UTC, comment #3:


Mistake: I meant this to go to bug #52722.
Will copy it there.

Michael Godfrey <godfrey>
Project Member
Sat 23 Dec 2017 12:14:24 AM UTC, comment #2:

Fix works for me.

Thanks and I will try to think of a way of testing
this kind of functionality.

All suggestions welcome, of course.


Michael Godfrey <godfrey>
Project Member
Fri 22 Dec 2017 11:58:04 PM UTC, comment #1:

OK, I see this problem and I can confirm that it is due to some kind of problem with (re)initialization of persistent variables as you mentioned in bug #52722. I'll try to take a look at that tomorrow. It is probably a similar problem it does not appear to have exactly the same cause.

John W. Eaton <jwe>
Project Administrator
Tue 19 Dec 2017 09:28:06 PM UTC, original submission:

ginput appears broken on dev. It looks like something is possibly resetting the persistent variables in the ginput_accumulator function. Sample code

The clicks are detected, as I found by putting a keyboard statement in the windowbuttondownfcn. But these somehow aren't signaling to the accumulator to stop.

Rik <rik5>
Project Administrator


(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):

No files currently attached


Depends on the following items: None found

   bug dependencies.


Carbon-Copy List
  • -unavailable- added by godfrey (Posted a comment)
  • -unavailable- added by jwe (Posted a comment)
  • -unavailable- added by rik5 (Submitted the item)

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only project members can vote.


    Please enter the title of George Orwell's famous dystopian book (it's a date):



    Follow 4 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Sun 24 Dec 2017 01:03:23 AM UTCrik5StatusReady For Test=>Fixed
    Sat 23 Dec 2017 07:35:16 PM UTCjweStatusNone=>Ready For Test
    Thu 21 Dec 2017 11:00:29 PM UTCrik5Dependencies-=>bugs #52693 is dependent

    Back to the top

    Powered by Savane 3.1-cleanup1