Changelog for package interactive_markers

1.12.0 (2020-04-26)

[feature] InteractiveMarkerServer::erase(): return false for unknown marker (#43)
[fix] Avoid overriding warning status messages (Fixes #34)
[fix] Fixed the problem that *.hpp was not installed (#66)
[fix] Make the Timer daemon survive /clock jumping back in a playback loop (#47)
[maintanence] Modernize package.xml and add missing std_msgs in CATKIN_DEPENDS
[maintanence] Use setuptools instead of distutils (#67)
[maintanence] Bump cmake version to 3.0.2 (#64)
[maintanence] Windows compatibility * Use C++11 portable sleep (#63) * Symbol visibility for shared libraries (#62) * Fix install location on Windows (#39)
Contributors: Alejandro Hernández Cordero, David Gossow, Fabian Maas, Jacob Perron, Robert Haschke, Sean Yen, Shane Loretz, shela

Fixed a crash when updates arrive, or are being processed, while shutdown is called (#36)
Contributors: Simon Schmeisser

The processFeedback function of the menu handler no longer catches the KeyErrors of the feedback_cb. See: #29
Added the empty() and size() members to InteractiveMarkerServer interface. See: #30
Contributors: Blake Anderson, Guglielmo Gemignani

Added explicit keyword argument queue_size for publisher in Python code and use the same default queue_size value as C++.
Fixed a SEGFAULT in setPose reported in #18 Previously, calling setPose() on an interactive marker causes a SEGFAULT if applyChanges() was not called on the server at least once since the marker was created. I traced the actual SEGFAULT to the doSetPose function. The value of header passed from setPose() is invalid because, in this case, marker_context_it = marker_contexts_.end(). I added a check for this case and, if there is no marker is present, instead use the header from the pending update.
Contributors: David Gossow, Mike Koval, William Woodall, ipa-fxm

remove debug statement that could produce segfault; init_it->msg->markers may be empty
Contributors: Filip Jares