I believe the only way to get OpenModelica to try to link against this library is if the configure script actually detects it is available.
What was the output of ./configure (./config.status -recheck to run it again).

Mine says: Simulations may use IPOPT: No, which means it will not try to link against it (ipopt is not used by the default solvers). I believe probably recent libipopt do not use libcoinmumps any longer but the configure script has not been updated.

Actually, looking at the configure script again, it only looks for the headers but does not verify that the libraries can be linked.
You can configure --without-ipopt to avoid using the library.

Thank you.

I configure it with "./configure --prefix=/usr/local --with-omniORB --with-ipopt --without-paradiseo" , because I have installed coinor-libipopt1 debian package
and all mumps packages in debian whheezy.
I'm going to try it by your reply I would like to create a script that would compile OpenModelica and created a deb package for wheezy, only with the support packages that are in wheezy.

Rony

Hi,

I have already re-compiled OpenModelica. I configure it with "./configure --prefix=/usr/local --with-omniORB --without-paradiseo", compilation with gcc-4.4 and ++-4.4, CFLAGS= -g -O2 -f PIC -ffloat-store, CXXFLAGS=-g -O2 -ffloat-store. I created a new RLC circuit to simulate the circuit in which the capacitor connected.
I compile this model with -ffloat-store and wihout -ffloat-store cflag, but no change. Compilation of this model lasted for 40 minutes and still not completed.
Any ideas please? How can I debug this?

Thank  you

Rony

So compilation takes that long? Which process is running after that time, omc or gcc?

OMEdit - 47 minutes, It was always OMEdit

OMEdit does not do any actual compilation work though. It's all sent to omc. If ps,top, or htop reports 47 minutes of CPU-time for OMEdit, you probably found a nasty error in OMEdit (but check if it's not omc or gcc taking time to be sure).
In that case I would have liked to see if the model works from command-line (omc a.mos).
a.mos:

loadModel(Modelica);getErrorString();

// more libraries maybe?

loadFile("myModel.mo");getErrorString();

translateModel(MyModelName);getErrorString();

This produces MyModelName.makefile. If this works, omc is fine. If make -f MyModelName.makefile spins, it's gcc at fault.

Also check if compilation actually worked, but the simulation is spinning (will be named something like Path.To.MyModelName).

omc from commandline create epc4 executable file. If I execute this file, this end after 1 second. This is a simulation end? Without errors.
So the error in OMEdit. How to find it?

Thank  you for help

Rony

Yes that should mean it finished successfully. You can check if $? is 0, then it is a success and epc4_res.mat is created. If then OMEdit starts spinning it is an OMEdit bug. I believe in the latest versions of OMEdit it is possible to open result files.
Try File->Open Result File and see if OMEdit locks up with just the result-file. It would then be good if you could then open a ticket using https://trac.openmodelica.org/OpenModelica/newticket and attach the result-file so the OMEdit maintainer can have a look at it.  It would also be good if the model could be provided if it is not confidential. If you do not want the files public, you can also email the files to adeel <dot> asghar <at> liu <dot> se.

These are exactly my symptoms:

https://www.openmodelica.org/index.php/ … pic?id=809

If I load result file to OMEdit, this plot it correctly without any errors. If I simulate from OMEdit, than above symptoms from this link.

Rony