Feat/add oscc interface node #1

rebpdx · 2017-09-05T18:53:39Z

This pull request adds a ROS node which exposes the OSCC API in the form of ROS messages so that a ROS user may utilize OSCC in their node environment.

To build and run tests:
Make sure you have Kinetic installed: http://wiki.ros.org/kinetic/Installation/Ubuntu

mkdir -p catkin_ws/src
cd catkin_ws/src
git clone --recursive https://github.com/PolySync/roscco.git
git checkout feat/add-oscc-interface-node
cd ..
catkin_make -DKIA_SOUL=ON
catkin_make run_tests -DKIA_SOUL=ON

Prior to this commit this repository was bare. This commit adds the initial interface for OSCC and ROS called roscco.

Prior to this commit the readme was lacking in information to a new user. This commit fixes that by providing information about what ROSCCO is, how to build it examples of using it and how to run the tests.

Prior to this commit the CMakefile and package.xml files had commented out example code provided from the catkin_create which is not needed to run this package. This commit fixes that by removing the commented out lines.

ZackPierce · 2017-09-07T02:12:27Z

Jenkinsfile

+      echo 'Release Package Created!'
+    }
+  }
+  catch(Exception e) {


If you're just going to rethrow the exception, there's no point in having the catch block here.

Nice catch. I was a little skeptic about groovy allowing the catch to be skipped but you would be correct, it is most certainly allowed and exercised.

ZackPierce · 2017-09-07T02:24:27Z

test/rapidcheck/include/rapidcheck.h

+#undef min
+#undef max
+
+#include "rapidcheck/Seq.h"


Pardon my unfamiliarity with the local conventions, but why are we including all of rapidcheck's source in this git repo as opposed to linking a submodule? Seems like unnecessary noise in the source tree and a possible cause of maintenance-cost-inducing drift.

I'd agree the submodule option is far better. I was having difficulty with building it in the catkin_make but apparently was a bit dense that day as it works as a submodule now.

Prior to this commit this repo used a static library and the header files for rapid check rather than a git submodule to work with ROS build system. This commit removes the duplicate headers and static library in favor of the submodule option while still working with the ros test process.

Prior to this commit there was a try catch finally block where the catch rethrew the same error so that the finally would still clean resources despite failure. This commit does away with the extra catch since it's possible to use a try finally in groovy.

Prior to this commit other projects depending on roscco would fail due to the roscco target library not being an artifact despite being listed as an output target. This commit fixes that by listing the actual output libraries.

Prior to this commit the rosco_node file had inconsistencies with formatting. This commit fixes that by updating the formatting to follow the rest of the code.

Prior to this commit ROS initialization would intermittently throw ROS errors not connecting to master but recover due to SIGIO interrupts from OSCC. This commit fixes that by blocking interupts during the ROS initialization process.

Prior to this commit this repository did not follow the ROS style guide. This commit changes that by using the clang-format file provied in the ROS style guide instrucitons. http://wiki.ros.org/CppStyleGuide

Prior to this commit there were catkin linter errors. This commit fixes that by addressing the errors produced.

Prior to this commit ROSCCO used an older version of OSCC this commit updates the dependency.

Prior to this commit ROSCCO did not disable before closing the OSCC connection relying on the 200 ms timoeout. This commit calls the disable before close to avoid relying on the timeout upon close.

Prior to this commit the topics advertised were using CamelCase rather than snake_case which is the ROS standard. This commit fixes that by changing the advertised topics to snake_case.

Prior to this commit this repo did not have an example about how to use ROSCCO. This commit fixes that by adding an example node that translates the ROS joystick node messages into ROSCCO messages with documantation about how the example works.

Prior to this commit there was a missing delete in one of the message callbacks. This commit fixes that by adding the required delete function to prevent memory leaks.

Prior to this commit the controller models were not called out and the steering did not have any smoothing algorithm making it difficult to control. This commit fixes that by calling out the game controllers used and adds an exponential average to the steering control.

Prior to this commit it was required to include defines in CMakeLists.txt for each available OSCC vehicle. OSCC now has a config file so that it does not need to be tracked in each project depending on OSCC. This commit utilizes that change.

Prior to this commit the joystick commander node example had non descript constants and variables. This commit fixes that by placing the values in named variables and adding comments to clarify what they are for and why.

nathanaschbacher · 2017-09-25T19:09:21Z

README.md

@@ -1 +1,81 @@
-# roscco


I'd prefer that the document structure be setup like...

Project Name

Overview

Stuff about what the point of this project is and what it's about

Getting Started

Dependencies

Building <whatever color-added text you'd like here>

Usage

Examples

API (if one exists)

Tests

Test Dependencies

Running Tests

License

Ye ol' MIT

The two most important parts of any open source project are the top-level headings... why it exists and what's its level of share-ittude.

nathanaschbacher · 2017-09-25T19:13:51Z

README.md

+
+
+## Building ROSCCO in a catkin project


I'm fairly certain that from this doc I wouldn't be able to set this thing up and actually run it on my machine... because I'm not sure where to start. I'm not sure what assumed knowledge or environment I'm supposed to already have, but it's such that running the build steps below on my machine produce some familiar errors.

So we're going to have to figure out how much is safe to assume about our user such that we can either disclaimer, "The following assumes that you already..." or provide more explicit setup instructions.

At least, as-is, I can't actually check this out and test it to approve the PR or not.

The same is true for the testing side.

nathanaschbacher · 2017-09-25T19:17:15Z

example/roscco_commander.cpp

+  enable_disable_pub_ = nh_.advertise<roscco::EnableDisable>("enable_disable", queue_size);
+
+  // Timed callback for publishing to OSCC < 200 ms
+  timer_ = nh_.createTimer(ros::Duration(0.05), &TeleopRoscco::timerCallback, this);


There are a variety of what are essentially constants throughout the code that should probably be hoisted. Like 0.05 and down on line #99, 0.99, and there are probably some more.

Also, unless they're supposed to be mutable we should set them up as consts whenever possible to ensure they're not mutated accidentally anywhere.

nathanaschbacher · 2017-09-25T19:19:13Z

package.xml

+<package>
+  <name>roscco</name>
+  <version>0.0.1</version>
+  <description>A OSCC to ROS interface</description>


nathanaschbacher · 2017-09-25T19:28:56Z

example/roscco_commander.cpp

+
+int main(int argc, char** argv)
+{
+  ros::init(argc, argv, "teleop_turtle");


teleop_turtle?

I was reading the turtle joystick conversion node example as a baseline to start from for following ROS standards on how the community handles joystick messages.

Actually there is more lacking consistency when comparing to the next line teleop_roscco and the node name roscco_commander. Since ROS community uses teleop for Teleoperation I'll make sure the node name and respective variables reflect roscco_teleop for clarity and consistency.

nathanaschbacher · 2017-09-25T19:30:02Z

I could also use some architectural overview for how the interaction of timers/timeouts is supposed to work.

Prior to this commit the trigger threshold was a mutable value which could yield undesirable results if the value was changed. This commit fixes that by setting the variable to be a constant.

Prior to this commit ROSCCO depended on a development branch version of OSCC rather than a released version of OSCC. This commit updates ROSCCO to depend on the released version 1.1.0. It also updates the Fault Report message to support the added dtcs value provided by the version upgrade.

Prior to this commit the documentation was lacking in clarity of what dependencies were required prior to using the components in ROSCCO. Also the README was lacking in clear header organizatoin. This commit fixes that by reorganizing the README headers, providing more details about the dependencies for this project and providing clarity of the Licensing information as well as corrections to the grammer in the description of the package.

Prior to this commit the README documentation used old variable naming for the example node. This commit fixes that by using the correct variable naming for the ROSCCO example.

Prior to this commit the example changes made it into a previous commit but were incomplete and contained dead code. This commit fixes that by removing dead code and using the correct variable names.

Prior to this commit the mapping functions that were added for clarity were mapped inversly to what their behavior are. This commit fixes that by mapping the triggers and steering to the desired behavior.

Prior to this commit the example deviated from the ROS coding standards. This commit fixes that by using the clang format to make it follow the ROS standard.

jlamb-at-polysync · 2017-10-17T19:23:52Z

example/roscco_teleop.cpp

+  const double data_smoothing_factor = 0.1;
+  double steering_average_;
+
+  // Variable to ensure joystick triggers have been initialized_


Is the extra _ character at the end intentional?

jlamb-at-polysync · 2017-10-17T19:29:34Z

example/roscco_teleop.cpp

+  ros::Subscriber joy_sub_;
+  ros::Timer timer_;
+
+  int previous_start_state_;


Is the initial value significant for these variables?

jlamb-at-polysync · 2017-10-17T19:48:55Z

msg/FaultOriginId.msg

@@ -0,0 +1,3 @@
+int32 FaultOriginBrake=0
+int32 FaultOriginSteering=1
+int32 FaultOriginThrottle=2


Is it a ROS convention to have enumerated types and alike be signed integers instead of unsigned?

jlamb-at-polysync · 2017-10-17T20:02:40Z

README.md

+```
+
+Similarly to the OSCC API The Reports and CanFrame messages can be


Should there be a comma after API?

jlamb-at-polysync · 2017-10-17T20:04:14Z

README.md

+
+Similarly to the OSCC API The Reports and CanFrame messages can be
+subscribed to as feedback about what is going on in OSCC. The Command and


Maybe this could read subscribed to in order to provide feedback?

jlamb-at-polysync · 2017-10-17T20:06:50Z

README.md

+EnableDisable messages will call the commands for the OSCC API.
+
+OSCC requires that a message to each device is sent within a 200ms, ROSCCO


Maybe this could be:
OSCC modules will disable if no messages are received within 200 ms. It is recommended to publish messages to the OSCC modules at a rate of 20 Hz (every 50 ms).

How about, "OSCC modules will disable if no messages are received within 200 ms. It is recommended to publish ROSCCO messages, at a rate of 20 Hz (every 50 ms) to ensure OSCC modules do not disable." I would like to make sure there isn't confusion that it's the person publishing the ROSCCO messages who need to be aware of this.

jlamb-at-polysync · 2017-10-17T20:18:16Z

msg/CanFrameData.msg

@@ -0,0 +1,6 @@
+uint32 can_id
+uint8 can_dlc
+uint8 pad


Do we need these reserved/internal fields mapped to the SocketCAN can_frame structure definition (they don't exist on all platforms and are for internal usage typically)?

jlamb-at-polysync · 2017-10-17T20:19:06Z

example/roscco_teleop.cpp

+}
+
+// Calculate the exponential average
+double calc_exponential_average(double average, double setpoint, double factor)


Could these function parameters be constant?

jlamb-at-polysync · 2017-10-17T20:21:06Z

example/roscco_teleop.cpp

+
+  // Smooth the steering to remove twitchy joystick movements
+  const double data_smoothing_factor = 0.1;
+  double steering_average_;


Is the initial value significant for this variable?

jlamb-at-polysync · 2017-10-17T20:22:20Z

example/roscco_teleop.cpp

+  // Number of messages to retain when the message queue is full
+  const int queue_size = 10;
+
+  // Timed callback frequency set to meet OSCC < 200 ms messages requirement


Might be more clear to say OSCC recommends publishing at a rate of 20 Hz (50 ms == 0.05 s)

Prior to this commit the CanFrame message contained reserved values which may or may not exist on other targets. This commit fixes that by removing them and assigning the values on an individual variable basis.

Prior to this commit some variables were in an initializer list while other were initialized in the class definition also there weren't const infront of function arguments. This commit clarifies the example by placing const infront of the function arguments and places all initilization values with the variable decelaration.

Prior to this commit there was some confusing vocabulary. This commit fixes that by restructuring some of the sentences.

Prior to this commit the FaultOriginId was a signed int for the enum which is conflicting with the OSCC enum type. This commit fixes that by using unsigned int.

jlamb-at-polysync · 2017-10-18T13:54:12Z

msg/CanFrameData.msg

@@ -1,6 +1,3 @@
 uint32 can_id
 uint8 can_dlc
-uint8 pad
-uint8 res0
-uint8 res1
 uint8[8] data


Is it possible to also define a constant for this size in the message definition file, just so you don't have to know its 8 bytes? SocketCAN uses CAN_MAX_DLEN for example in /usr/include/linux/can.h.

Something like this is what I had in mind:
const uint8 CAN_FRAME_DATA_MAX_SIZE = 8
...
uint8[CAN_FRAME_DATA_MAX_SIZE] data
...
for(int i = 0; i < CAN_FRAME_DATA_MAX_SIZE; i += 1)

Prior to this commit it was assumed the data message array size was known. This commit fixes that by adding a constant to the CanFrameData msg file so that it is clear what size the data array is.

Prior to this commit the repo was missing documentation for the public functions and the constants did not have consistant naming. This commit fixes that by using noisy snake case on constant variables across files and adds doxygen comments to all public functions.

vertexmachina · 2017-11-01T21:47:54Z

include/roscco/ros_to_oscc.h

+  /**
+   * @brief RosToOscc class initializer
+   *
+   * This functoin constructs ROS subscribers which can publish messages to OSCC API.


Typo: functoin

vertexmachina · 2017-11-01T21:49:10Z

include/roscco/oscc_to_ros.h

+  /**
+   * @brief Cast OSCC report to ROS message and publish
+   *
+   * This function casts an oscc report type to a ros message and publishes the message onto ROS. Where the ROSMSGTYPE


This function description block has inconsistent capitalization of ROS and OSCC.

vertexmachina · 2017-11-01T21:50:31Z

include/roscco/ros_to_oscc.h

+  /**
+   * @brief Callback function to publish ROS SteeringCommand messages to OSCC.
+   *
+   * This function is a callback that consumes a ROS BrakeCommand message and publishes them to the OSCC API.


Should be ROS SteeringCommand

Prior to this commit this repo had spelling errors and lower case acronyms. This commit fixes that by correcting the spelling errors and making the acronyms uppercase.

ZackPierce

👍

ZackPierce · 2017-10-27T16:10:58Z

src/oscc_to_ros.cpp

+    ROS_ERROR("Failed to block SIGIO");
+  }
+
+  topic_brake_report_ = public_nh->advertise<roscco::BrakeReport>("brake_report", 10);


Should probably use an explicatory constant, e.g. queue_size here, as below.

ZackPierce · 2017-10-27T16:14:53Z

src/ros_to_oscc.cpp

+  }
+
+  topic_brake_command_ =
+      public_nh->subscribe<roscco::BrakeCommand>("brake_command", 10, &RosToOscc::brakeCommandCallback, this);


Same story re: explaining arbitrary constant for queue_size.

ZackPierce · 2017-10-27T16:17:05Z

example/roscco_teleop.cpp

+  int previous_back_state_ = 0;
+
+  // Number of messages to retain when the message queue is full
+  const int queue_size = 10;


The ROS style guide seems to suggest SHOUTY_SNAKE case for constants. http://wiki.ros.org/CppStyleGuide#Constants

ZackPierce · 2017-10-27T16:24:01Z

include/roscco/oscc_to_ros.h

+  template <class OSCCTYPE, class ROSTYPE, class ROSSUBTYPE>
+  static void cast_callback(OSCCTYPE* report, ros::Publisher* pub);
+
+  static void steering_callback(oscc_steering_report_s* report);


Distinctly lacking in commentary for public functions. Not a blocking concern for the current PR, but may be worth considering in the future.

rebpdx added 3 commits September 1, 2017 16:14

Add initial OSCC and ROS interface

606ae04

Prior to this commit this repository was bare. This commit adds the initial interface for OSCC and ROS called roscco.

Update readme

4c325cc

Prior to this commit the readme was lacking in information to a new user. This commit fixes that by providing information about what ROSCCO is, how to build it examples of using it and how to run the tests.

Remove commented out code

2e302be

Prior to this commit the CMakefile and package.xml files had commented out example code provided from the catkin_create which is not needed to run this package. This commit fixes that by removing the commented out lines.

rebpdx requested review from nathanaschbacher and ZackPierce September 5, 2017 18:53

ZackPierce reviewed Sep 7, 2017

View reviewed changes

rebpdx added 15 commits September 7, 2017 12:25

Remove excess catch block

3b3aafd

Prior to this commit there was a try catch finally block where the catch rethrew the same error so that the finally would still clean resources despite failure. This commit does away with the extra catch since it's possible to use a try finally in groovy.

Fix CMakeLists target output

2ea0ec2

Prior to this commit other projects depending on roscco would fail due to the roscco target library not being an artifact despite being listed as an output target. This commit fixes that by listing the actual output libraries.

Format roscco_node for consistancy

f3a11d8

Prior to this commit the rosco_node file had inconsistencies with formatting. This commit fixes that by updating the formatting to follow the rest of the code.

Fix ROS connect to master errors

2ed22de

Prior to this commit ROS initialization would intermittently throw ROS errors not connecting to master but recover due to SIGIO interrupts from OSCC. This commit fixes that by blocking interupts during the ROS initialization process.

Format according to ROS style guide

f3c293b

Prior to this commit this repository did not follow the ROS style guide. This commit changes that by using the clang-format file provied in the ROS style guide instrucitons. http://wiki.ros.org/CppStyleGuide

Fix catkin_lint errors

775efa9

Prior to this commit there were catkin linter errors. This commit fixes that by addressing the errors produced.

Update OSCC dependency

97435bc

Prior to this commit ROSCCO used an older version of OSCC this commit updates the dependency.

Disable before closing OSCC connection

748274f

Prior to this commit ROSCCO did not disable before closing the OSCC connection relying on the 200 ms timoeout. This commit calls the disable before close to avoid relying on the timeout upon close.

Update topics to follow ROS standard

4ec1a52

Prior to this commit the topics advertised were using CamelCase rather than snake_case which is the ROS standard. This commit fixes that by changing the advertised topics to snake_case.

Add roscco_commander example

9f67714

Prior to this commit this repo did not have an example about how to use ROSCCO. This commit fixes that by adding an example node that translates the ROS joystick node messages into ROSCCO messages with documantation about how the example works.

Add missing delete

e467aec

Prior to this commit there was a missing delete in one of the message callbacks. This commit fixes that by adding the required delete function to prevent memory leaks.

Update CMakeList to use OSCC config file

51115b6

Prior to this commit it was required to include defines in CMakeLists.txt for each available OSCC vehicle. OSCC now has a config file so that it does not need to be tracked in each project depending on OSCC. This commit utilizes that change.

Clarify constants and variables

02ea756

Prior to this commit the joystick commander node example had non descript constants and variables. This commit fixes that by placing the values in named variables and adding comments to clarify what they are for and why.

nathanaschbacher reviewed Sep 25, 2017

View reviewed changes

package.xml Outdated

<package>

<name>roscco</name>

<version>0.0.1</version>

<description>A OSCC to ROS interface</description>

Copy link

nathanaschbacher Sep 25, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

An

nathanaschbacher reviewed Sep 25, 2017

View reviewed changes

rebpdx added 3 commits September 25, 2017 14:44

Make game trigger threshold mutable

6242183

Prior to this commit the trigger threshold was a mutable value which could yield undesirable results if the value was changed. This commit fixes that by setting the variable to be a constant.

rebpdx added 4 commits September 26, 2017 13:38

Update the README docunmentation for the example

982886c

Prior to this commit the README documentation used old variable naming for the example node. This commit fixes that by using the correct variable naming for the ROSCCO example.

Fix variable naming and remove dead code

9a531da

Prior to this commit the example changes made it into a previous commit but were incomplete and contained dead code. This commit fixes that by removing dead code and using the correct variable names.

Map triggers correclty

259fe94

Prior to this commit the mapping functions that were added for clarity were mapped inversly to what their behavior are. This commit fixes that by mapping the triggers and steering to the desired behavior.

Format example to ROS coding standards

a937847

Prior to this commit the example deviated from the ROS coding standards. This commit fixes that by using the clang format to make it follow the ROS standard.

jlamb-at-polysync reviewed Oct 17, 2017

View reviewed changes

README.md Outdated

```

Similarly to the OSCC API The Reports and CanFrame messages can be

Copy link

jlamb-at-polysync Oct 17, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should there be a comma after API?

jlamb-at-polysync reviewed Oct 17, 2017

View reviewed changes

rebpdx added 4 commits October 17, 2017 15:30

Remove socketcan reserved values

81aa12f

Prior to this commit the CanFrame message contained reserved values which may or may not exist on other targets. This commit fixes that by removing them and assigning the values on an individual variable basis.

Clarify README

e338b5b

Prior to this commit there was some confusing vocabulary. This commit fixes that by restructuring some of the sentences.

Use unsigned integers for enums

c9f8be7

Prior to this commit the FaultOriginId was a signed int for the enum which is conflicting with the OSCC enum type. This commit fixes that by using unsigned int.

jlamb-at-polysync reviewed Oct 18, 2017

View reviewed changes

Add CAN frame size constant to the data message

1a47782

Prior to this commit it was assumed the data message array size was known. This commit fixes that by adding a constant to the CanFrameData msg file so that it is clear what size the data array is.

jlamb-at-polysync approved these changes Oct 24, 2017

View reviewed changes

jlamb-at-polysync requested a review from vertexmachina October 27, 2017 15:32

vertexmachina suggested changes Nov 1, 2017

View reviewed changes

cotos added 2 commits November 2, 2017 12:43

Correct spelling errors and capitalization

10bd09c

Prior to this commit this repo had spelling errors and lower case acronyms. This commit fixes that by correcting the spelling errors and making the acronyms uppercase.

Correct miss acronym capitalization

2a69cd5

vertexmachina approved these changes Nov 2, 2017

View reviewed changes

cotos merged commit 2a69cd5 into devel Nov 2, 2017

ZackPierce reviewed Nov 3, 2017

View reviewed changes

rebpdx deleted the feat/add-oscc-interface-node branch November 3, 2017 23:38

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feat/add oscc interface node #1

Feat/add oscc interface node #1

rebpdx commented Sep 5, 2017 •

edited

Loading

ZackPierce Sep 7, 2017

rebpdx Sep 7, 2017

ZackPierce Sep 7, 2017

rebpdx Sep 7, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

nathanaschbacher Sep 25, 2017

rebpdx Sep 25, 2017 •

edited

Loading

nathanaschbacher commented Sep 25, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017 •

edited

Loading

rebpdx Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 17, 2017

jlamb-at-polysync Oct 18, 2017 •

edited

Loading

vertexmachina Nov 1, 2017

vertexmachina Nov 1, 2017

vertexmachina Nov 1, 2017

ZackPierce left a comment

ZackPierce Oct 27, 2017

ZackPierce Oct 27, 2017

ZackPierce Oct 27, 2017

ZackPierce Oct 27, 2017

		```

		Similarly to the OSCC API The Reports and CanFrame messages can be


		Similarly to the OSCC API The Reports and CanFrame messages can be
		subscribed to as feedback about what is going on in OSCC. The Command and

		EnableDisable messages will call the commands for the OSCC API.

		OSCC requires that a message to each device is sent within a 200ms, ROSCCO

Feat/add oscc interface node #1

Feat/add oscc interface node #1

Conversation

rebpdx commented Sep 5, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Project Name

Overview

Getting Started

Dependencies

Building <whatever color-added text you'd like here>

Usage

Examples

API (if one exists)

Tests

Test Dependencies

Running Tests

License

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rebpdx Sep 25, 2017 • edited Loading

Choose a reason for hiding this comment

nathanaschbacher commented Sep 25, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jlamb-at-polysync Oct 17, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jlamb-at-polysync Oct 18, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ZackPierce left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rebpdx commented Sep 5, 2017 •

edited

Loading

rebpdx Sep 25, 2017 •

edited

Loading

jlamb-at-polysync Oct 17, 2017 •

edited

Loading

jlamb-at-polysync Oct 18, 2017 •

edited

Loading