#18 Derivatives from window

[Maurizio De Pitta']

There is a serious bug apparently. It seems impossible to specify the derivatives through the option "from window" upon definition of a new system. I had issues with my system (see Discussion topics here) and then tried again with a tested example from the manual such as the bratu.m (also attached). I specified the symbolic derivatives through the window. No error is produced in the command line window. However when I open my odefile, there is no specification for the Jacobian. And also when I ask to edit my system by the GUI, the derivative widnow show me empty fields. The attached files show the system after I created. I hope this can be fixed.

[hilmeijer]

So this bug (or better non-functionality) does not affect 99.9% of users.
I confirm that I can reproduce this behaviour with a simple testcase too.

For now, either not use this functionality, OR better implement it manually after the system m-file has been created.

To be continued...

---

From: Maurizio De Pitta' mauriziodepitta@users.sourceforge.net
Sent: Saturday, May 2, 2020 2:09 PM
To: Ticket #18: Derivatives from window
Subject: [matcont:bugs] #18 Derivatives from window

---

** [bugs:#18] Derivatives from window**

Status: open
Group: v1.0_(example)
Labels: symbolic jacobian
Created: Sat May 02, 2020 12:09 PM UTC by Maurizio De Pitta'
Last Updated: Sat May 02, 2020 12:09 PM UTC
Owner: hilmeijer
Attachments:

• bratugefland.zip (sourceforge.net) (11.3 kB; application/zip)

There is a serious bug apparently. It seems impossible to specify the derivatives through the option "from window" upon definition of a new system. I had issues with my system (see Discussion topics here) and then tried again with a tested example from the manual such as the bratu.m (also attached). I specified the symbolic derivatives through the window. No error is produced in the command line window. However when I open my odefile, there is no specification for the Jacobian. And also when I ask to edit my system by the GUI, the derivative widnow show me empty fields. The attached files show the system after I created. I hope this can be fixed.

---

Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/matcont/bugs/18/

To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/

[Maurizio De Pitta']

Thanks.
Yes indeed, I was jeopardized because when I originally tried it, it was not working -- and the reason for it was a SEGFAULT in my mex code. The current way, using the GUI to define the system's ODE and then manually change the odefile to include a user-supplied Jacobian works. A remark for whoever wants to try it: besides the definition of the @jacobian function passed to handle(3) of the output of the odefile, do not miss to include an opt=odeset('Jacobian',handles(3)); in the init function of the odefile. (Attached an updated bratu.m example).