[WIP] Add WeightChangeDetector to tell weight change without candidates

[pazeshun]

Add new node weight_change_detector. This comes from the detection part of weight change from reset in weight_candidates_refiner. This can run without getting candidates.

cf. https://github.com/start-jsk/jsk_apc/pull/2659#issuecomment-419643335

不特定の物体を扱うためlabel_names.yamlやobject_weight.yamlがない状況で、電子天秤を物体を持ち上げたかどうかの判定のみに使いたくて(changed_from_resetトピックの利用)、candidates_publisherがいなくても動くようにしました。

[pazeshun]

On kinetic test,

Errors << jsk_interactive_marker:make /home/travis/ros/ws_jsk_apc/logs/jsk_interactive_marker/build.make.000.log

/opt/ros/kinetic/lib/libjsk_topic_tools.so: undefined reference to `operator<<(std::ostream&, XmlRpc::XmlRpcValue const&)'

collect2: error: ld returned 1 exit status

make[2]: *** [/home/travis/ros/ws_jsk_apc/devel/.private/jsk_interactive_marker/lib/jsk_interactive_marker/marker_6dof] Error 1

make[1]: *** [CMakeFiles/marker_6dof.dir/all] Error 2

make: *** [all] Error 2

[pazeshun]

In travis test, something wrong with baxter model generation on kinetic:

;; Input file is: baxter.dae

;; Config file is: baxter.yaml

;; Output file is: baxter.l

;; robot_name is: baxter

larm@7

rarm@15

head@5

QH6013 qhull input error: input is less than 3-dimensional since it has the same x coordinate

While executing:  | qhull C-0.001

Options selected for Qhull 2015.2 2016/01/18:

  run-id 1959395013  Centrum-premerge- 0.001  _max-width 0.04

  Error-roundoff 7.4e-17  _one-merge 0.003  _near-inside 0.015

  Visible-distance 0.001  U-coplanar-distance 0.001  Width-outside 0.002

  _wide-facet 0.006

;; points 72

[jsk_2016_01_baxter_apc:make] PQP Error! EndModel() called on model with no triangles

[jsk_arc2017_baxter:make] PQP Error! EndModel() called on model with no triangles

Perhaps related to: https://github.com/jsk-ros-pkg/jsk_model_tools/pull/216

[wkentaro]

This is a suggestion. For fast development, what about put them on the private repo for now, and merge it to jsk_apc later?

[pazeshun]

For fast development, what about put them on the private repo for now, and merge it to jsk_apc later?

I'm taking this approach now. Thank you.

[pazeshun]

Errors in https://github.com/start-jsk/jsk_apc/pull/2660#issuecomment-419687884 are also occurred in my local environment, but test passed and loading Euslisp model succeeded. So those errors are not currently fatal.

From travis log:

[jsk_2016_01_baxter_apc:make] Call Stack (max depth: 20):

[jsk_2016_01_baxter_apc:make]   0: at 152937952

[jsk_2016_01_baxter_apc:make]   1: at 152938048

[jsk_2016_01_baxter_apc:make]   2: at 105975136

[jsk_2016_01_baxter_apc:make]   3: at 105974872

[jsk_2016_01_baxter_apc:make]   4: at 105974032

[jsk_2016_01_baxter_apc:make]   5: at 105974032

[jsk_2016_01_baxter_apc:make]   6: at 106127384

[jsk_2016_01_baxter_apc:make]   7: at 106122488

[jsk_2016_01_baxter_apc:make]   8: at 106046160

[jsk_2016_01_baxter_apc:make]   9: at 106045752

[jsk_2016_01_baxter_apc:make]   10: at 115484960

[jsk_2016_01_baxter_apc:make]   11: at 115484960

[jsk_2016_01_baxter_apc:make]   12: at 115484960

[jsk_2016_01_baxter_apc:make]   13: at 117787760

[jsk_2016_01_baxter_apc:make]   14: at 115484600

[jsk_2016_01_baxter_apc:make]   15: at 117782864

[jsk_2016_01_baxter_apc:make]   16: at 115483688

[jsk_2016_01_baxter_apc:make]   17: at 115483688

[jsk_2016_01_baxter_apc:make]   18: at 115483688

[jsk_2016_01_baxter_apc:make]   19: at 110704584

[jsk_2016_01_baxter_apc:make]   And more...

[jsk_arc2017_baxter:make] Call Stack (max depth: 20):

[jsk_arc2017_baxter:make]   0: at 119260864

[jsk_arc2017_baxter:make]   1: at 119260960

[jsk_arc2017_baxter:make]   2: at 112284408

[jsk_arc2017_baxter:make]   3: at 112284096

[jsk_arc2017_baxter:make]   4: at 112283832

[jsk_arc2017_baxter:make]   5: at 112283832

[jsk_arc2017_baxter:make]   6: at 112280328

[jsk_arc2017_baxter:make]   7: at 112273488

[jsk_arc2017_baxter:make]   8: at 112302480

[jsk_arc2017_baxter:make]   9: at 112302408

[jsk_arc2017_baxter:make]   10: at 126772040

[jsk_arc2017_baxter:make]   11: at 126772040

[jsk_arc2017_baxter:make]   12: at 126772040

[jsk_arc2017_baxter:make]   13: at 110074352

[jsk_arc2017_baxter:make]   14: at 126771680

[jsk_arc2017_baxter:make]   15: at 110069864

[jsk_arc2017_baxter:make]   16: at 126467472

[jsk_arc2017_baxter:make]   17: at 126467472

[jsk_arc2017_baxter:make]   18: at 126467472

[jsk_arc2017_baxter:make]   19: at 127703000

[jsk_arc2017_baxter:make]   And more...

[jsk_arc2017_baxter:make] 84798520 unittest-error: 116095736 84928800 in 119260864, exitting...

[jsk_arc2017_baxter:make] [ERROR] test 119260864 failed ... (116095736 12).

[jsk_arc2017_baxter:make] Call Stack (max depth: 20):

[jsk_arc2017_baxter:make]   0: at 146672552

[jsk_arc2017_baxter:make]   1: at 146672648

[jsk_arc2017_baxter:make]   2: at 134243064

[jsk_arc2017_baxter:make]   3: at 134242752

[jsk_arc2017_baxter:make]   4: at 134242488

[jsk_arc2017_baxter:make]   5: at 134242488

[jsk_arc2017_baxter:make]   6: at 134238984

[jsk_arc2017_baxter:make]   7: at 134232144

[jsk_arc2017_baxter:make]   8: at 134261136

[jsk_arc2017_baxter:make]   9: at 134261064

[jsk_arc2017_baxter:make]   10: at 148740152

[jsk_arc2017_baxter:make]   11: at 148740152

[jsk_arc2017_baxter:make]   12: at 148740152

[jsk_arc2017_baxter:make]   13: at 132033008

[jsk_arc2017_baxter:make]   14: at 148739792

[jsk_arc2017_baxter:make]   15: at 132028520

[jsk_arc2017_baxter:make]   16: at 148738832

[jsk_arc2017_baxter:make]   17: at 148738832

[jsk_arc2017_baxter:make]   18: at 148738832

[jsk_arc2017_baxter:make]   19: at 149661656

[jsk_arc2017_baxter:make]   And more...

[jsk_arc2017_baxter:make] 106757176 unittest-error: 145059080 106887456 in 146672552, exitting...

[jsk_arc2017_baxter:make] [ERROR] test 146672552 failed ... (145059080 12).

[jsk_arc2017_baxter:make] [Testcase: testtest_pick_interface] ... FAILURE!

[jsk_arc2017_baxter:make] [Testcase: testtest_stow_interface] ... FAILURE!

Strange point 1: call stack outputted number instead of code Strange point 2: tests in jsk_2016_01_baxter_apc passed though call stack appeared Strange point 3: test names were number

[pazeshun]

Moved to #2667 and #2668