[Octave-bug-tracker] [bug #60208] (optim) lsqnonlin error when outputfcn terminates algorithm

[anonymous]

URL:
  <https://savannah.gnu.org/bugs/?60208>

                 Summary: (optim) lsqnonlin error when outputfcn terminates
algorithm
                 Project: GNU Octave
            Submitted by: None
            Submitted on: Wed 10 Mar 2021 03:58:39 PM UTC
                Category: Octave Forge Package
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Unexpected Error or Warning
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: jasmin.zimmer@batop.de
             Open/Closed: Open
                 Release: 6.2.0
         Discussion Lock: Any
        Operating System: Any

    _______________________________________________________

Details:

Hello, 

I've added an output function to stop lsqnonlin for a specific condition. For
stop=true I'm getting the error 


error: rmfield: structure does not contain field lambda
error: called from
    lsqnonlin at line 282 column 10


for _some_ inputs. I've looked into the outp variable and sometimes it's
simply missing the "lambda" field. 
One of my outputs from lsqnonlin is also the jacobian; in the stop=true case,
same thing happens in lines 290/291 because residmin_out{4} doesn't contain a
"lambda" field. 

This happens in both Octave 6.2 and an older dev branch version (May 2020),
with optim package versions 1.6.0 and 1.6.1. 

I'm pretty sure this is an actual bug, the question is just whether/how it
depends on the actual input/optimisation values. I haven't included a minimal
example because I'm not quite sure how to replicate the bug without attaching
half of our other project code... 




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?60208>

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