bugGNU Octave - Bugs: bug #54995, Removal of called_from_builtin

 
 

bug #54995: Removal of called_from_builtin

Submitter:  Guillaume <gyom>
Submitted:  Fri 09 Nov 2018 04:56:17 PM UTC
 
Category:  Interpreter Severity:  5 - Blocker
Priority:  5 - Normal Item Group:  Regression
Status:  Fixed Assigned to:  None
Originator Name:  Guillaume Open/Closed:  * Closed
Release:  * dev Operating System:  * Any
* Mandatory Fields

Add a New Comment (Rich Markup)
   

Comment Type & Canned Response:
       

 

( Jump to the original submission )

Wed 16 Jan 2019 08:12:20 PM UTC, comment #11: 

With the test fixed, closing as fixed as per comment #7.

Markus Mützel <mmuetzel>
Project Member
Tue 15 Jan 2019 07:13:45 PM UTC, comment #10: 

@Guillaume: I renamed the class like you proposed and no longer see the regression.
Pushed to stable here:
http://hg.savannah.gnu.org/hgweb/octave/rev/0447e44cd18b

Markus Mützel <mmuetzel>
Project Member
Mon 14 Jan 2019 11:19:14 AM UTC, comment #9: 

It works if I run the test manually but there is indeed an error with 'make check'. The full error message from the log file is:

***** test <*54995>
 obj = testclass ();
 assert (obj.x(1), 1);
!!!!! regression: https://octave.org/testfailure/?54995
mismatch in number of fields

The issue seems to be that there is another test with a class with the same name, bug-36025/@testclass/testclass.m, that is executed beforehand in the same Octave session. Maybe we should just rename testclass to be testclass54995?

Guillaume <gyom>
Sun 13 Jan 2019 11:29:49 AM UTC, comment #8: 

The new test is failing on the buildbots:

>>>>> processing /scratch/buildbot/workers/jwe-debian-x86_64-0/stable-clang-4_0-debian/src/test/bug-54995/bug-54995.tst
***** test <*54995>
 obj = testclass ();
 assert (obj.x(1), 1);

Markus Mützel <mmuetzel>
Project Member
Sat 12 Jan 2019 04:50:22 PM UTC, comment #7: 

Thank you; in my eyes, it was indeed a blocker but it's now all back to normal.

Guillaume <gyom>
Sat 12 Jan 2019 02:18:56 PM UTC, comment #6: 

I backed out the change and added the test from comment #2 on stable and merged those changes with default.

John W. Eaton <jwe>
Project Administrator
Sat 12 Jan 2019 07:48:28 AM UTC, comment #5: 

I forgot to mention that Dynare (http://www.dynare.org) would become basically unusable if this issue is not fixed.

Sébastien Villemot <svillemot>
Sat 12 Jan 2019 07:46:04 AM UTC, comment #4: 

For me it is a serious regression. And it's easy to fix (just reverting one commit).

Sébastien Villemot <svillemot>
Sat 12 Jan 2019 12:02:29 AM UTC, comment #3: 

@jwe: Is this a blocker for the 5.0 release?

Rik <rik5>
Project Administrator
Wed 21 Nov 2018 10:17:06 AM UTC, comment #2: 

I attach a test case corresponding to the example below.  It is a regression for a common construct when writing @classes so I would classify its severity as high.

Writing a minimal test case for bug #32296 will take longer but I think fixing it is lower priority than the one mentioned here.

(file #45505)

Guillaume <gyom>
Fri 09 Nov 2018 05:32:26 PM UTC, comment #1: 

When I removed called_from_builtin, I looked for a bug report or message about it but didn't find anything.

We should add both of these test cases with bug numbers to the build system for future reference.  I can do that but it may take a few days.

John W. Eaton <jwe>
Project Administrator
Fri 09 Nov 2018 04:56:17 PM UTC, original submission:  

The called_from_builtin function was recently removed:

https://hg.savannah.gnu.org/hgweb/octave/rev/920a4ad8043b

It has the consequence of reintroducing the bug described here:

http://lists.gnu.org/archive/html/help-octave/2009-04/msg00307.html

and initially fixed there (where called_from_builtin was added):

http://hg.savannah.gnu.org/hgweb/octave/rev/b2b8ed43b922

For simplicity, I copy here a small example that illustrates the bug:

# Save as @myclass/myclass.m
function obj = myclass ()

  obj = struct ("a", rand(1));
  obj = class(obj, "myclass");

endfunction

# Save as @myclass/subsref.m
function val = subsref (obj, S)

  idx.type = S.type;
  idx.subs = {S.subs};

  val = builtin ("subsref", obj, idx);

endfunction

octave> obj = myclass ();
octave> obj.a
error: max_recursion_depth exceeded
error: called from
    subsref at line 6 column 7

A positive consequence is that it fixes bug #32296, e.g. Mike's example from comment #7 now works.

Guillaume <gyom>

 

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

Attach Files:
   
   
Comment:
   

Attached Files
file #45505:  bug-54995.zip added by gyom (2KiB - application/zip)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by mmuetzel (Posted a comment)
  • -email is unavailable- added by svillemot (Posted a comment)
  • -email is unavailable- added by rik5 (Posted a comment)
  • -email is unavailable- added by svillemot
  • -email is unavailable- added by jwe (Posted a comment)
  • -email is unavailable- added by gyom (Submitted the item)
  •  

    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 project members can vote.

     

     

     

     

    Follow 6 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2019-01-16 mmuetzel StatusReady For Test Fixed
        Open/ClosedOpen Closed
    2019-01-12 jwe Severity3 - Normal 5 - Blocker
        StatusNone Ready For Test
    2018-12-12 svillemot Carbon-Copy- Added svillemot
    2018-11-21 gyom Attached File- Added bug-54995.zip, #45505

    Back to the top


    Powered by Savane 3.9