OMEdit crashes for each PowerSystems model

[Rüdiger Franke]

This bug seems to be quite new in the current trunk version. The crash can (hopefully) be reproduced as follows:

$ git clone https://github.com/modelica/PowerSystems.git
$ OMEdit PowerSystems/PowerSystems/package.mo

In the left Libraries pane:

• open PowerSystems.Examples.Network
• double click on NetworkLoop

Select the menu Simulation -> Check Model

[Martin Sjölund]

Check of PowerSystems.Examples.Network.NetworkLoop completed successfully.
Class PowerSystems.Examples.Network.NetworkLoop has 198 equation(s) and 198 variable(s).
75 of these are trivial equation(s).

No errors.

[Rüdiger Franke]

Here is the content of my openmodelica.stacktrace.OMEdit:

Caught signal 11 (SIGSEGV)
../OpenModelica/build/bin/OMEdit(_Z13signalHandleri+0x29f)[0x4ad46f]
/lib/x86_64-linux-gnu/libc.so.6(+0x350e0)[0x7fe2534dd0e0]
/usr/lib/x86_64-linux-gnu/libstdc++.so.6(__dynamic_cast+0x19)[0x7fe253dc7609]
../OpenModelica/build/bin/OMEdit(_ZN11ModelWidget12validateTextEPP15LibraryTreeItem+0x29)[0x57af69]
../OpenModelica/build/bin/OMEdit(_ZN10MainWindow10checkModelEP15LibraryTreeItem+0x3a)[0x4c175a]
../OpenModelica/build/bin/OMEdit(_ZN10MainWindow10checkModelEv+0x32)[0x4c4df2]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN11QMetaObject8activateEP7QObjectPKS_iPPv+0x32c)[0x7fe25442d71c]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN7QAction9triggeredEb+0x42)[0x7fe254c92ce2]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN7QAction8activateENS_11ActionEventE+0x87)[0x7fe254c94867]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(+0x599ab3)[0x7fe25506bab3]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN15QAbstractButton17mouseReleaseEventEP11QMouseEvent+0x74)[0x7fe25506bc04]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN11QToolButton17mouseReleaseEventEP11QMouseEvent+0xa)[0x7fe25512dd0a]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN7QWidget5eventEP6QEvent+0x298)[0x7fe254cec748]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN19QApplicationPrivate13notify_helperEP7QObjectP6QEvent+0x8c)[0x7fe254c9948c]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN12QApplication6notifyEP7QObjectP6QEvent+0x41f)[0x7fe254ca010f]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN16QCoreApplication14notifyInternalEP7QObjectP6QEvent+0x8d)[0x7fe25441871d]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN19QApplicationPrivate14sendMouseEventEP7QWidgetP11QMouseEventS1_S1_PS1_R8QPointerIS0_Eb+0x15f)[0x7fe254c9f76f]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(+0x244432)[0x7fe254d16432]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN12QApplication15x11ProcessEventEP7_XEvent+0x60c)[0x7fe254d14e2c]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(+0x26bed2)[0x7fe254d3ded2]
/lib/x86_64-linux-gnu/libglib-2.0.so.0(g_main_context_dispatch+0x24d)[0x7fe24fe0fc5d]
/lib/x86_64-linux-gnu/libglib-2.0.so.0(+0x49f48)[0x7fe24fe0ff48]
/lib/x86_64-linux-gnu/libglib-2.0.so.0(g_main_context_iteration+0x2c)[0x7fe24fe0fffc]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN20QEventDispatcherGlib13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE+0x5d)[0x7fe254447d1d]
/usr/lib/x86_64-linux-gnu/libQtGui.so.4(+0x26bf96)[0x7fe254d3df96]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN10QEventLoop13processEventsE6QFlagsINS_17ProcessEventsFlagEE+0x41)[0x7fe254417271]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN10QEventLoop4execE6QFlagsINS_17ProcessEventsFlagEE+0x1a5)[0x7fe2544175d5]
/usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN16QCoreApplication4execEv+0x99)[0x7fe25441d059]
../OpenModelica/build/bin/OMEdit(main+0x1045)[0x4a8d95]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7fe2534c9b45]
../OpenModelica/build/bin/OMEdit[0x4ad035]

Any idea?

[Rüdiger Franke]

Btw. I get the same crash for Modelica.Blocks.Examples.PID_Controller. Check model of Modelica.Blocks.Examples.Filter gives:

[1] 17:19:50 Syntax Warning
[: 1:59-1:61]: Treating .2 as 0.2. This is not standard Modelica and only done for compatibility with old code. Support for this feature may be removed in the future.

[2] 17:19:50 Syntax Error
[: 1:0-1:0]: No viable alternative near token: /

[Rüdiger Franke]

It works again after complete recompilation of OMEdit ...

[Rüdiger Franke]

OK, the status "fixed" wasn't correct indeed. The status "invalid" is also incorrect. Obviously OMEdit is missing a dependency on omc sources. The workaround was to rebuild OMEdit completely.

[Martin Sjölund]

No, you got a crash inside the C++ sources. Qmake does not fully support incremental builds and you need to make clean if there are bigger changes to the sources. I strongly believe it is not a problem with dependencies to OM. I used to get the same kind of errors even before there was any direct linking to OM sources.