Skip to content

Latest commit

 

History

History
493 lines (362 loc) · 22 KB

CONTRIBUTING.md

File metadata and controls

493 lines (362 loc) · 22 KB
Raw

Contributing to Eclipse iceoryx

Thanks for your interest in this project.

Project description

In domains like automotive, robotics or gaming, a huge amount of data must be transferred between different parts of the system. If these parts are actually different processes on a POSIX based operating system like Linux, this huge amount of data has to be transferred via an inter-process-communication (IPC) mechanism. Find more infos on the Eclipse site.

Eclipse Contributor Agreement

Before your contribution can be accepted by the project team, contributors must electronically sign the Eclipse Contributor Agreement (ECA).

Commits that are provided by non-committers must have a Signed-off-by field in the footer indicating that the author is aware of the terms by which the contribution has been provided to the project. The non-committer must additionally have an Eclipse Foundation account and must have a signed Eclipse Contributor Agreement (ECA) on file.

For more information, please see the Eclipse Committer Handbook.

Contact

Contact the project developers via the project's "dev" list.

Feature request and bugs

We love pull requests! The next sections try to cover most of the relevant questions. For larger contributions or architectural changes, we'd kindly ask you to either:

or

  • Create a design document and raise it in a separate pull request beforehand

If you would like to report a bug or propose a new feature, please raise an issue before raising a pull request. Please have a quick search upfront if a similar issue already exists. A release board is used to prioritize the issues for a specific release. This makes it easier to track the work-in-progress. If you have troubles getting an issue assigned to you please contact the maintainers via Gitter.

Please make sure you have:

  1. Signed the Eclipse Contributor Agreement
  2. Created an issue before creating a branch, e.g. Super duper feature with issue number 123
  3. All branches have the following naming format: iox-[issue]-branch-name e.g. iox-123-super-duper-feature
  4. All commits have the following naming format: iox-#[issue] Commit message e.g. iox-#123 Implement super-duper feature
  5. All commits have been signed with git commit -s
  6. The iceoryx-unreleased.md in doc/website/release-notes is updated with the GitHub issue that is closed by the Pull-Request
  7. You open your pull request towards the base branch main
  8. Link the pull request to the according GitHub issue and set the label accordingly

NOTE: For support while developing you can use little helper scripts, see git-hooks.

Experimental features

Large features or features where the API is not yet clear can be implemented with the IOX_EXPERIMENTAL_POSH feature flag. Those features shall not be available in the installed headers when the IOX_EXPERIMENTAL_POSH feature flag was not set during compilation.

If possible, this should be achieved by not installing the headers of the experimental feature instead of ifdefs. With this approach, the experimental features can still be build on all targets on the CI without specifying the IOX_EXPERIMENTAL_POSH feature flag and does not require new CI targets.

The experimental feature must be in an experimental namespace and the header includes path must also contain the name experimental.

The experimental features should not be mentioned in a prominent place so as not to encourage users to use them since there will be no community support and they might be removed at a later stage.

Branching strategy

main

  • Main development branch
  • Open for external contributions

release_x.y

  • Branch for stabilising a certain release
  • Write access limited to maintainers
  • Fine-tuning of external contribution e.g. running Axivion SCA
  • Finish any missing implementations regarding the quality levels

As depicted below, after the release branch has been created the stabilization phase will begin. After finishing the release, a git tag will be created to point to HEAD of the release branch. Follow-up releases will be branched off from the git tag.

o---o---o---o---o  main
     \
      \      v1.0.0      v1.0.1
       \        |           |
        o---o---o---o---o---o---o  release_1.0
                         \
                          \      v1.1.0
                           \        |
                            o---o---o  release_1.1

Coding style

We love the C++ core guidelines. If in doubt please try to follow them as well as our unwritten conventions in the existing parts of the code base. Please format your code with the provided clang-format and clang-tidy before raising a pull request. Nowadays, many IDEs read the clang-format file.

We created some handy rules to highlight some specifics that you might not be used to in other FOSS projects. They are helpful to build embedded systems for safety fields like automotive or avionics. It is possible that not the whole codebase follows these rules, things are work in progress.

  1. No heap is allowed, static memory management hugely decreases the complexity of your software (e.g. cxx::vector without heap)
  2. No exceptions are allowed, all function and methods need to have noexcept in their signature
  3. No undefined behavior, zero-cost abstract is not feasible in high safety environments
  4. Use C++17
  5. Rule of Five, if there is a non-default destructor needed, the rule of five has to be applied
  6. Keep the STL dependencies to a minimum, the building blocks in iceoryx_hoofs aim to be compatible with the STL, but the code may contain additions which are not compatible with the STL (e.g. iox::cxx::vector::emplace_back() does return a bool); see section below
  7. Always use iox::log::Logger, instead of printf()
  8. Always use iox::ErrorHandler or IOX_EXPECTS/IOX_ENSURES, when an error occurs that cannot or shall not be propagated via an iox::expected
  9. Not more than two-level nested namespaces, three-level nested namespace can be used sparsely

See error-handling.md for additional information about logging and error handling.

For formatting and linting rules on Bazel files see the installation guide for contributors.

Naming conventions

  • File names with lower_snake_case: my_thing.hpp
  • Structs, classes and enum classes in UpperCamelCase: class MyClass{}
  • Methods and variables in lowerCamelCase: uint16_t myVariable
  • Compile time constants, also enum values in UPPER_SNAKE_CASE: static constexpr uint16_t MY_CONSTANT
  • Class members start with m_: m_myMember
    • Public members of structs and classes do not have the m_ prefix
  • Namespaces in lower_snake_case : my_namespace
  • Aliases have a _t postfix : using FooString_t = iox::string<100>;
  • Objects created from a method returning a iox::optional<Foo> shall be named maybeFoo
  • Objects created from a method returning a iox::expected<Foo, FooError> shall contain the name result e.g. getChunkResult for the method getChunk()

clang-tidy suppressions

WARNING: never suppress concurrency-mt-unsafe! This warning can be emitted by free c functions. When such a function is used in a class one is not allowed to create multiple instances of such a class and access them from different threads. Even when every instance is contained in only one thread, the underlying thread-unsafe function is maybe accessed concurrently which can result in race conditions.

If required constructs create clang-tidy warnings one can suppress them with a justification and either NOLINTNEXTLINE(warning-type) or NOLINTBEGIN(warning-type) & NOLINTEND(warning-type).

Those suppressions always require a justification which has to be provided with NOLINTJUSTIFICATION. But do not repeat yourself in the justification. If the doxygen documentation, or a suppression in the header for the same declaration provides already an argument please refer to it. A justification should always state why the suppressed construct is required and how the code ensures the safe usage.

Furthermore, NOLINTBEGIN should only be used for a range as small as possible and maybe for a whole function but not more.

Examples

// NOLINTJUSTIFICATION we require the 'construct' to implement XXX and the safe usage
//                         is guaranteed through YYY
// NOLINTBEGIN(some-warning)
auto a = myLineOfCodeWithWarning();

Doxygen

Please use doxygen to document your code.

The following doxygen comments are required for public API headers:

/// @brief short description
/// @param[in] / [out] / [in,out] name description
/// @return description

For overrides of virtual methods the copydoc tag can be used:

/// @copydoc BaseClass::method
/// @note Optional describe some specifics to the override

A good example for code formatting and doxygen structure can be found in swe_docu_guidelines.md (WIP)

External dependencies

External dependencies such as the STL or other libaries shall be kept to a minium for iceoryx_posh and iceoryx_hoofs. If you think a new dependency is necessary, do the following:

  1. Contact the maintainers beforehand by opening an issue to discuss the necessity
  2. If accepted, add the new header to tools/scripts/used-headers.txt for the CI to pass

Eliminating code duplication

  1. In some cases, the code in the constructor and assignment operations may be largely duplicated. It is encouraged to move this duplicated code into a separate function (e.g., copy_and_move_impl) for better reuse. Additionally, the MoveAndCopyHelper in iceoryx/design (refer to MoveAndCopyHelper) offers some functionalities that make this process easier.

Folder structure

The folder structure boils down to:

  • iceoryx_COMPONENT
    • cmake: all CMake files go here, needed for find_pkg()
    • doc: manuals and documentation
    • include: public headers with stable API
      • internal: public headers with internal API, which might change quite frequently
    • source: implementation files
    • test: unit and integration tests
    • CMakeLists.txt: build the component separately
  • examples_iceoryx: Examples described in iceoryx_examples

All new code should follow the folder structure.

How to add a new example

  1. Add the example in the "List of all examples"
  2. Create a new file in doc/website/examples/foobar.md and add it to doc/website/examples/.pages. This file shall only set the title and include the readme from /iceoryx_examples/foobar/README.md
  3. Add an add_subdirectory directive into iceoryx_meta/CMakeLists.txt in the if(EXAMPLES) section.
  4. Consider using geoffrey for syncing code in code blocks with the respective source files
  5. Add integration test (add as dependency to package.xml and write a launch_test for the example)
  6. Record an asciicast and embed image link

Testing

We use Google test for our unit and integration tests. We require compatibility with the version 1.10.0.

Have a look at our best practice guidelines for writing tests and installation guide for contributors for building them.

Unit tests (aka module tests)

Unit tests are black box tests that test the public interface of a class. They are required for all new code.

Each unit test case needs a unique identifier (UUID according to RFC 4122) in the form of:

::testing::Test::RecordProperty("TEST_ID", "12345678-9ab-cdef-fedc-1234567890ab");

UUID can be for example generated with Python or the command line tool:

import uuid
uuid.uuid4()
uuidgen -r

In rare cases you may want to exclude a GoogleTest case from the execution (e.g. sporadic failures). When doing that you need to add a macro call right after the Test ID:

::testing::Test::RecordProperty("TEST_ID", "12345678-9ab-cdef-fedc-1234567890ac");
GTEST_SKIP() << "@todo iox-#1234 Enable test once the API is supported";

A technical reason and a valid ticket number is needed to track the re-enabling (or removing) of the test.

Integration tests

Integration tests test the interaction of several classes. They are optional for new code.

Coverage Scan

To ensure that the provided test code covers the productive code you can do a coverage scan with gcov. The reporting is done with lcov and htmlgen. You will need to install the following packages:

sudo apt install lcov

In iceoryx we have multiple test levels for test coverage: unit, integration, component and all for all test levels together. You can create reports for these different test levels or for all tests. Coverage is done with gcc. The coverage scan applies to Quality level 3 and partly level 2 with branch coverage.

To generate a coverage report, iceoryx needs to be compiled with coverage flags and the tests need to be executed. You can do this with one command in the iceoryx folder like this:

./tools/iceoryx_build_test.sh clean build-all -c <testlevel>

Optionally, you can use the build-all option to get the coverage for extensions like the C-Binding. The -c flag indicates that you want to generate a coverage report and requires you to pass the test level. By default the test level is set to all.

./tools/iceoryx_build_test.sh clean debug build-all -c unit

NOTE iceoryx needs to be built as static library to work with gcov flags. The script does this automatically.

The flag -c unit is for generating only reports for unit tests. In the script tools/scripts/lcov_generate.sh, the initial scan, filtering and report generation is done automatically.

All reports are stored locally in build/lcov as html report (index.html). In GitHub, we are using Codecov for general reporting of the code coverage. Codecov gives a brief overview of the code coverage and also indicates in Pull-Requests if newly added code is not covered by tests. If you want to see detailed html reports for specific Pull-Requests or branches, you can check here.

Legal & Compliance

Safety & security

The iceoryx maintainers aim for ASIL-D compliance. The ISO26262 is a good read if you want to learn more about automotive safety. A nice introduction video was presented on CppCon 2019.

If you want to report a vulnerability, please use the Eclipse process.

Static code analysis

The iceoryx maintainers have a partnership with Axivion and use their Axivion Suite to run a static code analysis.

Github labels are used to group issues into the rulesets:

Ruleset name Github issue label Priority
Adaptive AUTOSAR C++14 AUTOSAR ⭐⭐⭐
SEI CERT C++ 2016 Coding Standard CERT ⭐⭐
MISRA C++ 2008 MISRA ⭐

If one of the rules is not followed, a rationale is added in the following manner:

Either with a comment in the same line:

*mynullptr = foo; // AXIVION Same Line Ruleset-A1.2.3 : Short description why

Or with a comment one line above:

// AXIVION Next Line Ruleset-A1.2.3 : Short description why
*mynullptr = foo;

It is also possible to suppress a rule for a complete construct:

// AXIVION Construct Ruleset-A1.2.3 : Short description why
class Foo
{
  void doSomething()
  {
    // Do something useful
  }
};

Header

Each source file needs to have this header:

// Copyright (c) [YEAR OF INITIAL CONTRIBUTION] - [YEAR LAST CONTRIBUTION] by [CONTRIBUTOR]. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// SPDX-License-Identifier: Apache-2.0

NOTE: The date is either a year or a range of years with the first and last years of the range separated by a dash. For example: "2004" (initial and last contribution in the same year) or "2000 - 2004". The first year is when the contents of the file were first created and the last year is when the contents were last modified. The years of contribution should be ordered in chronological order, thus the last date in the list should be the year of the most recent contribution. If there is a gap between contributions of one or more calendar years, use a comma to separate the disconnected contribution periods (e.g. "2000 - 2004, 2006").

Example:

// Copyright (c) 2019 - 2020, 2022 by Acme Corporation. All rights reserved.
// Copyright (c) 2020 - 2022 by Jane Doe <[email protected]>. All rights reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// SPDX-License-Identifier: Apache-2.0

NOTE: For scripts or CMake files you can use the respective comment syntax # for the header.

Quality levels

The CMake targets are developed according to the ROS quality levels. Despite developing some targets according to automotive standards like ISO26262, the code base standalone does NOT legitimize the usage in a safety-critical system. All requirements of a lower quality level are included in higher quality levels e.g. quality level 4 is included in quality level 3.

Quality level 5

This quality level is the default quality level. It is meant for examples and helper tools.

  • Derived from ROS quality level 5
    • Reviewed by two approvers
    • No compiler warnings
    • License and copyright statements available
    • No version policy required
    • No unit tests required

Quality level 4

  • Derived from ROS quality level 4
    • Basic unit tests are required
    • Builds and runs on Windows, MacOS, Linux and QNX

Quality level 3

  • Derived from ROS quality level 3
    • Doxygen and documentation required
    • Test specification required
    • Version policy required

Quality level 2

This quality level is meant for all targets that need tier 1 support in ROS 2.

Quality level 1

  • Derived from ROS quality level 1
    • Version policy for stable API and ABI required
    • ASPICE SWE.6 tests available
    • Performance tests and regression policy required
    • Static code analysis warnings in Axivion addressed
    • Enforcing the code style is required
    • Unit tests have full statement and branch coverage

Quality level 1+

This quality level goes beyond the ROS quality levels and contains extensions.

  • Code coverage according to MC/DC available

Training material recommended for contributors

  • Effective C++ by Scott Meyers
  • Unit Testing and the Arrange, Act and Assert (AAA) Pattern by Paulo Gomes
  • The C++ Standard Library by Nicolai M. Josuttis
  • Modern C++ Programming with Test-Driven Development: Code Better, Sleep Better by Jeff Langr
  • Modern C++ Design: Generic Programming and Design Patterns by Andrei Alexandrescu
  • Exceptional C++ by Herb Sutter
  • C++ Concurrency in Action by Anthony Williams