Github workflow for unit tests

RDK Services

RDK services are a set of JSON-RPC based RESTful services for accessing various set-top box components. RDK Services are managed and accessed through the Thunder framework. Thunder supports both HTTP and Websocket requests, making the services easily accessible to Lightning, Web, and native client applications.

View Latest Documentation

Contributing to RDKServices

License Requirements

Before RDK accepts your code into the project you must sign the RDK Contributor License Agreement (CLA).
Each new file should include the latest RDKM license header.
License for this project is included in the root directory and there shouldn't be any additional license file in any of the subfolders.

How to contribute?

Fork the repository, commit your changes, build and test it in at least one approved test platform/device.
To test it in a RDKV device, update SRC_URI and SRCREV in the rdkservices_git.bb recipe to point to your fork.
Submit your changes as a pull request to the latest sprint branch.
If more than one developer has to work on a particular feature, request for a dev branch to be created.

Pull request Checklist

When a pull request is submitted, blackduck, copyright and cla checks will automatically be triggered. Ensure these checks have passed (turned into green).
At least one reviewer needs to review and approve the pull request.
For tracking and release management purposes, each pull request and all the commits in the pull request shall include RDK ticket number(s) or Github issue number(s) and “reason for the change”.
Any pull request from Comcast developers should include a link to successful gerrit verification (in the comment section).
To verify your changeset in gerrit, submit a test gerrit change to rdkservices_git.bb with the SRC_URI and SRCREV pointing to your fork.
If the changes to RDKServices require any Thunder framework changes, the contributor has to plan for a limited regression testing (with the Thunder and RDKServices changes) before submitting the pull request.

Comcast CI/CD

RDKServices branches - sprint vs main (Specific to RDKV Builds)

Comcast gerrit sprint branch will point to rdkservices sprint branch.
1. For example 2103_sprint branch gerrit recipe will point to sprint/2103 branch in github.
Comcast gerrit stable2 branch will point to rdkservices main branch in github.

Sprint/stable2 git hash updates (Specific to RDKV Builds)

Git hash in rdkservices_git.bb (sprint_branch) will be updated periodically (at least once a week).
1. Dev owner to follow up as needed and update the JIRA ticket to “ready for sprint testing”.
Once a change is verified in the sprint branch and approved by RM, the developer can contact the maintainers to cherry-pick the change set to the main branch or submit the cherry picked change set to the main branch.
1. Git hash in rdkservices_git.bb (stable2) will be updated periodically.
2. Dev owner to follow up as needed and update the ticket to “ready for release testing”.
What if a changeset in the sprint branch fails sprint testing?
1. The developer has to submit a pull request to undo the commit before the end of the sprint cycle or
2. The changeset will be abandoned in the sprint branch and won’t make into the main/stable2

Upstream Vs Patch

Patches will increase the chances of build failures when the git hash is moved to a newer version.
We encourage everyone to upstream all the changes to GitHub instead of using patches
On a need basis, a developer can request the maintainers for an approval to use a patch in RDKV build (as a stop-gap measure). An unapproved patch will be rejected.

Documentation

RDK services are described using JSON Schema. JSON Schema provides a standard approach for describing APIs and ensures consistency across all APIs. There are two schemas that are used to describe a service:

plugin.schema.json: A schema for defining a service.
interface.schema.json: A schema for defining the properties, methods, and events of a service.

Each RDK service has an instance of these schemas in the root of the service folder. For example, MyServicePlugin.json and MyService.json. These files are used to generate API documentation as Markdown. Each service has a Markdown file that is written to the docs/api folder. The following demonstrates the folder structure:

/rdkservices
    /MyService
        /MyService.json
        /MyServicePlugin.json
    /docs/api
        /MyServicePlugin.md

Markdown files are generated from the JSON definitions using the json_generator tool (Tools/json_generator/generator_json.py).

The generator tool requires:

Python 3.5 or higher
The jsonref library

Verify your Python version:

python --version

Install jsonref if it is not already installed:

pip install jsonref

Generating Markdown for a Single Service

To generate markdown for a single service:

Change directories to Tools/json_generator.
Run generator_json.py and provide the location of the service JSON plugin file using the -d argument and the output directory using the -o argument. You must also include the --no-interfaces-section argument; otherwise, an interface section is added to the markdown that links back to the ThunderInterfaces project. Make certain that you are pointing to the plugin definition and not the interface definition. Here is an example of using the tool:
```
python ./generator_json.py -d ../../MyService/MyServicePlugin.json  -o ../docs/api --no-interfaces-section --verbose $files
```
The MyServicePlugin.md file is written to the ../docs/api folder. This is the standard directory where all the service API markdown files are written.

Generating Markdown for All Services

A script is provided to generate the markdown for all services and does a complete build of the documentation. The script only generates the markdown for a service if the JSON definition has been updated. In addition, the script post-processes the generated markdown files to create standard link anchors and to clean the build.

To generate markdown for all services:

From the rdkservices repository, change directories to docs/Tools/md_generator.
Run generate_md.py. For example:
```
python ./generate_md.py
```
All markdown files are written to the ../docs/api folder. This is the standard directory where all the service API markdown files are written.

Use the existing services as a guide when learning the structure of both the plugin and interface schemas.

Coding Guidelines

Style Guide

The point of having style guidelines is to have a common vocabulary of coding so people can concentrate on what you’re saying rather than on how you’re saying it. If the code you add to a file looks drastically different from the existing code around it, it throws readers out of their rhythm. Please Avoid this. If you’re editing code, take a few minutes to determine the coding style of the component and apply the same style.
- All RDK Services must have a callsign with a prefix of org.rdk. RDK Service name must use PascalCase (first letter of each word is capitalized). For instance org.rdk.PersistentStore
- API & Event Names should use camelCase (first word is lowercase & first letter of subsequent words are capitalized). For instance getDeviceInfo, onAppResumed.
- JSON parameter names in request and responses should use camelCase and be valid ascii strings. For instance "updateAvailable".
- API getters should start with get. For instance getDefaultInterface. Setters should start with set. For instance setDefaultInterface.
- Event names should follow the convention on[Noun Object][Action Verb].
  - Object should be a noun and provide useful context on the event.
  - Action should be a verb. Use present and past tenses to indicate if the event is triggered before or after.
  - For instance onAppResumed, onAppResuming.
- Capitalize acronyms by default. So use setHDMIEnabled instead of setHdmiEnabled.
- Ensure JSON-RPC API responses return either result or error object on success or error respectively. Please refer to JSON-RPC Spec for more details on expected members in the response object. Highlighting important members below.
  - result
    - This member is REQUIRED on success.
    - This member MUST NOT exist if there was an error invoking the method.
  - error
    - This member is REQUIRED on error.
    - This member MUST NOT exist if there was no error triggered during invocation.
    - The value for this member MUST be an Object as defined below.
      - code
        
        A Number that indicates the error type that occurred.
        
        This MUST be an integer.
      - message
        
        A String providing a short description of the error.
        
        The message SHOULD be limited to a concise single sentence.
  - Either the result member or error member MUST be included, but both members MUST NOT be included.
- To maintain uniformity in all text-editors, set TAB size to 2 or 4 spaces and replace TAB by SPACES.
RDK services are implemented as Thunder Plugins and must adhere to the PluginHost::IPlugin interface. If a RDK service handles WEB requests it implements PluginHost::IWeb. If it activates/deactivates and handles JSON-RPC it implements PluginHost::IDispatcher (or derives from PluginHost::JSONRPC). If it implements custom interfaces it adds them to ThunderInterfaces for RPC. If it exposes interfaces to other processes via RPC it develops an RPC::Communicator. A service specifies its interfaces like this:

    BEGIN_INTERFACE_MAP(MyPlugin)
    INTERFACE_ENTRY(PluginHost::IPlugin)
    INTERFACE_ENTRY(PluginHost::IDispatcher)
    END_INTERFACE_MAP

MODULE_NAME
- Thunder provides a trace and warning reporting feature. To accurately identify the source of a warning, Thunder needs to know the human readable name of the package (executable or library). This package name is defined by the MODULE_NAME and declared by the MODULE_NAME_DECLARATION()
- Any package that includes a Thunder component requires such a definition and declaration. If the definition is missing, a compiler error will be reported (error missing MODULE_NAME) and if the declaration is missing, a linker error will be reported (missing or duplicate symbol)
- Module.h defines the mandatory MODULE_NAME. It also includes framework headers (<plugins/plugins.h>)
```
#ifndef MODULE_NAME
#define MODULE_NAME Plugin_IOController
#endif
```
- Module.cpp defines the mandatory MODULE_NAME_DECLARATION
```
#include "Module.h"
MODULE_NAME_DECLARATION(BUILD_REFERENCE)
```
A service is registered in a translation unit via a mandatory SERVICE_REGISTRATION(MyService, API_VERSION_NUMBER_MAJOR, API_VERSION_NUMBER_MINOR, API_VERSION_NUMBER_PATCH)
MyPlugin.config can specify: autostart, startuporder, custom properties (passed to the service during activation via PluginHost::IShell::ConfigLine()). During the project configuration, write_config(MyPlugin) in CMakeLists.txt uses MyPlugin.config to generate and install a corresponding json file. Refer to PersistentStore.config
MyPlugin.json documents JSON-RPC interface. It's added to both plugin folder and ThunderInterfaces. The latter generates classes for parameters or enums that can be included like #include <interfaces/json/JsonData_MyPlugin.h>. In the plugin code, JSON-RPC methods and properties are registered like this:

    Register<void /*input params*/, void /*output params*/>(_T("sync"), &MyPlugin::endpoint_sync, this);
    Property<LocationData>(_T("location"), &MyPlugin::get_location /*getter*/, nullptr /*setter*/, this);

Initialization and Cleanup
- Keep Plugin Constructors & Destructors lean. Most initialization should be done within Initialize() which gets called when the plugin is activated. More on that below. This will ensure that WPEFramework boots up faster since most of the plugins are not auto-started or activated on bootup. Similarly most of the plugin cleanup should happen within Deinitialize() instead of the destructor.
- Prefer to do Plugin Initialization within IPlugin Initialize(). If there is any error in initialization return non-empty string with useful error information. This will ensure that plugin doesn't get activated and also return this error information to the caller. Ensure that any Initialization done within Initialize() gets cleaned up within IPlugin Deinitialize() which gets called when the plugin is deactivated.
- Ensure that any std::threads created are joined within Deinitialize() or the destructor to avoid std::terminate exception. Use the ThreadRAII class for creating threads which will ensure that the thread gets joined before destruction.
Inter-plugin communication - There might be use cases where one RDK Service or plugin needs to call APIs in another RDK Service. Don't use JSON-RPC for such communication since it's an overhead and not preferred for inter-plugin communication. JSON-RPC must be used only by applications. Instead use COM RPC through the IShell Interface API QueryInterfaceByCallsign() exposed for each Plugin. Here is an example.

Versioning

Versioning
- Given a version number MAJOR.MINOR.PATCH, increment the:
  - MAJOR version when you make incompatible API changes that break backwards compatibility. This could be removing existing APIs, changes to API Signature or major changes to API behavior that breaks API contract,
  - MINOR version when you add backward compatible new features like adding new APIs, adding new parameters to existing APIs,
  - PATCH version when you make backwards compatible bug fixes.
- RDK Service version in MAJOR.MINOR.PATCH format is updated using SERVICE_REGISTRATION macro.
```
SERVICE_REGISTRATION(DisplaySettings, API_VERSION_NUMBER_MAJOR, API_VERSION_NUMBER_MINOR, API_VERSION_NUMBER_PATCH);
```
- There is also a Plugin::Metadata structure maintained for each RDK Service that keeps the versioning information. This is returned in call to Controller.1.status.
```
static Plugin::Metadata<Plugin::DisplaySettings> metadata(
        // Version (Major, Minor, Patch)
        API_VERSION_NUMBER_MAJOR, API_VERSION_NUMBER_MINOR, API_VERSION_NUMBER_PATCH,
        // Preconditions
        {},
        // Terminations
        {},
        // Controls
        {}
    );
}
```
- Changes in version should be updated when commits are added to the main or release branches. There should be a version change per JIRA Ticket. This is not enforced on sprint branches since there could be multiple changes for the same JIRA ticket during development.
Changelog
- Each RDK Service has a CHANGELOG file that contains all changes done so far. When version is updated, add a entry in the CHANGELOG.md at the top with user friendly information on what was changed with the new version. Please don't mention JIRA tickets in CHANGELOG. Refer to Changelog as an example and Keep a Changelog for more details.
- Please Add entry in the CHANGELOG for each version change and indicate the type of change with these labels:
  - Added for new features.
  - Changed for changes in existing functionality.
  - Deprecated for soon-to-be removed features.
  - Removed for now removed features.
  - Fixed for any bug fixes.
  - Security in case of vulnerabilities.
- Changes in CHANGELOG should be updated when commits are added to the main or release branches. There should be one CHANGELOG entry per JIRA Ticket. This is not enforced on sprint branches since there could be multiple changes for the same JIRA ticket during development.
Deprecation
- Breaking changes to the API that requires a major version update should first go through Deprecation by doing a minor version update. We recommend atleast 2 RDK releases with the deprecated API/s and minor version update to give time for clients and apps to make changes to remove the deprecated API. Following needs to be done for deprecation.
  - The API/s getting deprecated should be marked with a "deprecated" label in the json schema. This will ensure that it's updated in the API documentation.
  - Add a changelog entry with minor version update and include Deprecated label to call out the API/s getting deprecated.
  - If this API/s is getting replaced by a newer API then it can come in the same minor version update with changelog entry with Added label.

Code Review Checklist

This checklist is primarily intended for maintainers or reviewers. Please check for the following before approving Pull Requests.

Coding Guidelines are followed.
API Changes are documented and versioned.
JIRA Ticket Number is mentioned in the title or commit message for tracking.
Any Dependent HAL changes are to be merged first before merging changes to RDK Service to avoid build issues.
For a New RDK Service, ensure autostart flag is set to false. The general recommendation is for RDK Services to not be autostarted or activated unless it's required on bootup. Resident Apps can activate the required RDK Services on demand.
Approve Pull Requests to main branch or release branches (release/*) only after you get Release Management approval for those specific branches.

Questions?

If you have any questions or concerns reach out to Anand Kandasamy

For a plugin specific question, maintainers might refer you to the plugin owner(s).

	#define getNumberParameter(paramName, param) { \
	if (WPEFramework::Core::JSON::Variant::type::NUMBER == parameters[paramName].Content()) \
	param = parameters[paramName].Number(); \
	else \
	try { param = std::stoi( parameters[paramName].String()); } \
	catch (...) { param = 0; } \
	}

	do{
	network_available = checkNetwork();
	if ( !network_available ){
	sleep(30);

	"summary": "Delay offset (in ms) on the selected audio port",
	"type": "string",
	"example": "0"

	"summary": "The over temperature grace interval",
	"type": "string",
	"example": "600"

	TEST_F(HdmiInputDsTest, writeEDID)
	{
	EXPECT_EQ(Core::ERROR_NONE, handler.Invoke(connection, _T("writeEDID"), _T("{\"deviceId\": 0, \"message\": \"message\"}"), response));
	EXPECT_EQ(response, string("{\"success\":true}"));
	}

	void HdmiInput::writeEDID(int deviceId, std::string message)
	{

	}

	//give Netflix process some time to terminate gracefully.
	sleep(10);

	notify(RDKSHELL_EVENT_ON_WILL_DESTROY, params);
	sleep(gWillDestroyEventWaitTime);

	killAllApps();

	do{
	if ((getServiceState(m_service, "org.rdk.AuthService", state) != Core::ERROR_NONE) \|\| (state != PluginHost::IShell::state::ACTIVATED)) {
	sleep(10);

	while(m_currentArcRoutingState != ARC_STATE_ARC_TERMINATED)
	{
	usleep(500000);
	}

	std::cout << "bounds set\n";
	usleep(68000);
	std::cout << "all set\n";

	"description": "User name is already taken (i.e. the user has already joined the room)",
	"$ref": "#/common/errors/illegalstate"

	while(tts_request.empty())
	{
	sleep(60);

	std::string script = "echo '" + path + "' \| sed -r \"s/([^$])([$\\{])([^$\\{\\}\\/])(.)/\\3/\"";
	variable = Utils::cRunScript(script.c_str());

	script = ". /etc/device.properties; echo \"$" + variable + "\"";
	value = Utils::cRunScript(script.c_str());

	const char * script1 = R"(grep DEVICE_TYPE /etc/device.properties \| cut -d "=" -f2 \| tr -d '\n')";
	m_isHybridDevice = Utils::cRunScript(script1).substr();

	command = "find "+path+" -iname "+filter+" > /tmp/tempBuffer.dat";
	retStat = system(command.c_str());

	std::string command = "mkdir -p " + dir + " \0";
	Utils::cRunScript(command.c_str());

	snprintf(cmd, 127, "rm -f %s", STANDBY_REASON_FILE);
	pipe = popen(cmd, "r");

rdkcentral / rdkservices Goto Github PK

rdkservices's Introduction

RDK Services

Table of Contents

Contributing to RDKServices

License Requirements

How to contribute?

Pull request Checklist

Comcast CI/CD

RDKServices branches - sprint vs main (Specific to RDKV Builds)

Sprint/stable2 git hash updates (Specific to RDKV Builds)

Upstream Vs Patch

Documentation

Generating Markdown for a Single Service

Generating Markdown for All Services

Coding Guidelines

Versioning

Code Review Checklist

Questions?

rdkservices's People

Stargazers

Watchers

Forkers

rdkservices's Issues

Recommend Projects

Recommend Topics

Recommend Org