bugGNU Octave - Bugs: bug #57709, Octave crashes when changing...

 
 

bug #57709: Octave crashes when changing sub-structure that is viewed in Variable Editor

Submitter:  Dennis <dizze>
Submitted:  Fri 31 Jan 2020 01:25:59 PM UTC
   
 
Category:  GUI Severity:  3 - Normal
Priority:  5 - Normal Item Group:  Segfault, Bus Error, etc.
Status:  Fixed Assigned to:  None
Originator Name:  Open/Closed:  * Closed
Release:  * 5.2.0 Operating System:  * Microsoft Windows
Fixed Release:  None Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Jump to the original submission

Wed 13 May 2020 10:22:24 AM UTC, comment #11: 

AFAICS all examples in this bug report work fine for me with both Octave-6.0.1 (built specifically to test JWE's cset over in bug #58317) and dev Octave-7.0.0 from yesterday.

Closing report with status "fixed".
The fix will be in the next Octave release; that was due any moment but got delayed because of unforeseen blocker issues. Hopefully it'll be released this spring.

@OP: if you want you can try that 6.0.1 (but you'll find that it is quite slow - one of the aforementioned blockers), see here:
https://prn183.stackstorage.com/s/wFhbA8rZXtPnAUu
and alternatively there's a 7.0.0 from May 11 as well, used here also:
https://prn183.stackstorage.com/s/YC9JvFeX3RZyfJW
Both links expire next Monday, May 18. Please don't disseminate them; they're not nightly builds but rather personal crossbuilds with several personal mods.

Rik, thanks for systematically stepping through the series of open VE bug reports. I've entered quite few of those and commented in several others but lost track, possibly because there was no feedback (could be I wasn't in the CC lists of the latter).

Philip Nienhuis <philipnienhuis>
Group Member
Tue 12 May 2020 09:58:41 PM UTC, comment #10: 

@Philip: Could you check whether the fix for bug #58317 by jwe also fixes this bug report?

Rik <rik5>
Group administrator
Sun 02 Feb 2020 09:07:21 PM UTC, comment #9: 

Philip,
I checked bug #55668 and there are similarities, but also differences. If I understand the comments in that thread correctly, the error there results from a VE action that is incorrectly translated into a code to perform the action. However, the bug I report here occurs when a valid code is executed in the command line while viewing the struct in the VE. The system doesn't crash when the code is executed without viewing the variable. It seems that the adjustment of the VE window after changing the struct goes wrong. That is not necessarily the same as what happens when editing a struct from within the VE.

Dennis <dizze>
Sun 02 Feb 2020 05:28:50 PM UTC, comment #8: 

@Dennis:
This bug looks a lot like bug #55668. Can you confirm?

Philip Nienhuis <philipnienhuis>
Group Member
Sat 01 Feb 2020 12:05:52 AM UTC, comment #7: 

@Dennis:
Thanks for the info.

Intriguing, this has been working in the development version until recently.
I'll try some more binary installers (I made one ~ every 3 days) to see where it started to crash.

A bit of experimenting shows that the VE doesn't like the size of a.b{1} to be made smaller:

a.b{1} = {4 5 6} ## works fine
a.b{1} = {7 8}  ## crashes
a.b{1} = {9 10 11 12} ## works fine
but then:
a.b{1} = {1 2 3} ## crash

Philip Nienhuis <philipnienhuis>
Group Member
Fri 31 Jan 2020 08:38:00 PM UTC, comment #6: 

Philip, I have seen this for at least a year, in all versions since v4.4.1.

(I didn't use Octave before that, but worked with Matlab).

Dennis <dizze>
Fri 31 Jan 2020 08:30:00 PM UTC, comment #5: 

@OP, Dennis:

_This bug is troubling me for a while now_

Since when do you see this happening?
I've tried a few older 6.0.0 Windows installers (development release) I have lying around and one from Jan 19 still works fine.

Philip Nienhuis <philipnienhuis>
Group Member
Fri 31 Jan 2020 04:23:03 PM UTC, comment #4: 

I don't have windows.
I did try the original part of bug #53858 and it does not crash  but does go into an infinite loop of error messages.

Doug Stewart <dastew>
Fri 31 Jan 2020 03:31:55 PM UTC, comment #3: 

Works for me too in Octave-5.1.0, but not in the dev version (from Jan. 26 and from Jan. 29) on Windows. So probably some sort of regression.

When leaving the semicolon at the end of 'a.b(1) = []' Octave does show the new contents of 'a' and returns to the prompt but segfaults very soon after. Hopefully that gives a clue.
Absence or presence of the 'clear all' statement makes no difference.

Philip Nienhuis <philipnienhuis>
Group Member
Fri 31 Jan 2020 03:31:37 PM UTC, comment #2: 

Thanks Doug. I have two questions:
1. do you have the possibility to confirm this in Windows?
2. Could you try bug #53858 in Ubuntu? The described behaviour is still present in v5.1.90.

Pending confirmation, it seems this bug is a Windows issue.

Dennis <dizze>
Fri 31 Jan 2020 02:07:23 PM UTC, comment #1: 

I tried what you said and did not see any problems. It worked as expected.
Ubuntu and octave 5.1.9
hgid
ans = ece72b94486f

Doug Stewart <dastew>
Fri 31 Jan 2020 01:25:59 PM UTC, original submission:  

This bug is troubling me for a while now and I hoped it would be solved in v.5.1.90. It is not. It is present in v4.4.1 and up.

The bug has a lot of resemblance with bug id #53858. However, the example provided there doesn't crash the GUI, but causes an endless loop. The example below actually causes a crash of the GUI

Method to reproduce:

clear all
a.b{1} = {1 2 3};
a.b{2} = {2 3};
a.b{3} = [];
openvar a;
% Double-click the  "a.b" cell to open a new Variable Editor panel
% Double-click the  "a.b{1}" cell to open a new Variable Editor panel
a.b{1} = [];

The last entry causes Octave to crash (tested in Windows).

Dennis <dizze>

 

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

Attach Files:
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by rik5 (Posted a comment)
  • -email is unavailable- added by dastew (Posted a comment)
  • -email is unavailable- added by dizze (Submitted the item)
  • -email is unavailable- added by dizze
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only group members can vote.

     

    Follow 8 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2020-05-13 philipnienhuis StatusReady For Test Fixed
        Open/ClosedOpen Closed
    2020-05-12 rik5 StatusConfirmed Ready For Test
    2020-01-31 rik5 Carbon-CopyRemoved 72865 -
    2020-01-31 rik5 Release5.1.90 5.2.0
    2020-01-31 philipnienhuis Item GroupNone Segfault, Bus Error, etc.
        StatusNone Confirmed
    2020-01-31 dizze Carbon-Copy- Added dizze

    Back to the top

    Powered by Savane 3.13-3230.
    Corresponding source code