cloud_merge crashing after sweeping

[pschillinger]

As seen today during the lamor workshop after doing a sweep:

[INFO] [WallTime: 1440691238.322840] Sweep action server started
[INFO] [WallTime: 1440691238.443560] ptu_pan_tilt_metric_map client created
[INFO] [WallTime: 1440691248.425980] Shortest sweep type: -140 60 160 -30 -30 -30
[ INFO] [1440691248.464078704]: Pan tilt sweep started -> acquire data
[ INFO] [1440691256.760380545]: Room observation now has 307200 points
[ INFO] [1440691260.972871361]: Room observation now has 614400 points
[ INFO] [1440691265.405918678]: Room observation now has 921600 points
[ INFO] [1440691269.669232021]: Room observation now has 1228800 points
[ INFO] [1440691274.171145983]: Room observation now has 1536000 points
[ INFO] [1440691278.571035152]: Room observation now has 1843200 points
[ INFO] [1440691283.663835428]: Pan tilt sweep stopped
[ INFO] [1440691283.687719434]: Created semantic map summary xml.
[ INFO] [1440691283.689908429]: Found 1 results.
ASSERT failure in QList<T>::operator[]: "index out of range", file /usr/include/qt4/QtCore/qlist.h, line 477
[cloud_merge-1] process has died [pid 6939, exit code -6, cmd /opt/ros/indigo/lib/cloud_merge/cloud_merge __name:=cloud_merge __log:=/home/phil/.ros/log/56c2a7ee-4cd1-11e5-98f4-f4b7e2032701/cloud_merge-1.log].
log file: /home/phil/.ros/log/56c2a7ee-4cd1-11e5-98f4-f4b7e2032701/cloud_merge-1*.log

We solved it by manually deleting the "metric_maps" DB in mongodb.