PyCharm debug segmentation error (signal 11)

Question

In PyCharm (community version 2016.2.3) using anaconda2 + ubuntu 14.04, importing matplotlib raises error 11 during debug mode. There are no problems in the release model at launch time.

python code:

import matplotlib as pt

debug console:

Connected to the pydev debugger (build 162.1967.10) GLib-GIO-Message: using GSettings "memory". Your settings will not be saved or transferred to other applications. Backend Qt4Agg is an interactive backend. Turn on interactive mode.

Process terminated by exit code 139 (interrupted by signal 11: SIGSEGV)

Answer 1

I had the same error message. I uninstalled anaconda and installed miniconda instead and reinstalled pycharm. Exit code 139 came when I ran the python console as well as the debug console.

Now, when I import matplotlib, I still get the messages:

Backend Qt4Agg is an interactive backend. Turn on interactive mode.

GLib-GIO message: GSettings memory usage. Your settings will not be saved or transferred to other applications.

I assume this is not perfect, but I am not getting an exit code, and everything seems to work. This is a solution, but there must be a more reliable way.

Answer 2

A bit late, but a googler might help someone.

Qt may cause this problem. PyCharm works with -qt-support = auto by default. If you have python bindings set for Qt4 and Qt5, the auto function may not select the correct version of Qt. Try setting the correct Qt bindings in PyCharm settings (Build, Ex ... → Python Debugger - compatible with PyQt)

Setting up from Auto to PyQt4 worked for me in the conda environment without deleting anything.

Answer 3

My work worked after I removed the pyqt5 bindings.

sudo apt-get remove python3-pyqt5

Answer 4

I managed to get rid of segfault and code 139 by disabling the Qt window in PyCharm settings (Build, Ex ... → Python Debugger). My use case: running code using a remote interpreter. You need to use matplotlib only for a specific shape.

Answer 5

In my case, this was due to the pandas package. He was probably doing dataframes in streaming (not supported ?!).

To find the reason, do the following:

 gdb python (gdb) script.py 

(and upon failure)

 where 

This will show the stack trace. In my case, the missing file in numpy that was fixed:

pip install --upgrade pandas