Coding Standards

Lots of people work together on Smoothie’s code.


layout: default title: Contributing to Smoothieware —

Contributing

Contributing

The Smoothieware project is always looking for help.

Whatever your skills are, there are things you can do to improve the project with other volunteers.

Help and working together is what has made this project so advanced and precious over the years, and is very welcome.

We appreciate all contributions, whether it's code, documentation, bug reports, or helping other users in the community!

Getting Started

Here is a list of the useful resources you want to read if you want to contribute to the project:

Ways to Contribute

Code Contributions

If you can code, there are always features to implement and bugs to fix.

Check the Todo list for current priorities.

Documentation

Help improve this documentation by fixing errors, adding examples, or writing new guides.

See Editing the Wiki for instructions.

Community Support

Answer questions on IRC, the forum, or mailing lists.

Helping other users is a valuable contribution!

Testing

Test new features on different hardware configurations and report your findings.

Beta testing helps catch issues before releases.

Contact

If you want to contribute but aren’t sure how, please contact us.

We’ll help you find something that matches your skills and interests!

In order to make everybody’s life easier, and keep the codebase clean and manageable, here are a few rules you should follow when coding Smoothie stuff:

Do not include other headers in a header file unless absolutely necessary to define something in that header

Un-necessary includes clutter the code, and makes compile time longer.

Bad example ( in a .h file ):

#include "Kernel.h"
#include "PanelScreen.h"
#include "Button.h"
#include <array>
#include <vector>

Instead do:

#include <array>
#include <vector>

class Kernel;
class PanelScreen;
class Button;

However, you still need to include whatever is necessary in your .cpp file.

Use class pointers in classes and use Class xxx in header to forward define it

Same thing: no unnecessary includes in .h files:

class Robot;
class Stepper;
class Planner;

class Kernel {
    public:
        Kernel();

        Robot* robot;
        Stepper* stepper;
        Planner* planner;
};

Include headers only in cpp files, and only ones that are needed to fully define a class

Never include something in the header file (unless absolutely necessary). Include it in the .cpp file, and only if it is really required.

Do not put code in a method in a header unless absolutely necessary

The only time this is needed is for time-critical code that must be inlined.

Do not use #ifdef in the code

This is rarely needed and is only currently used by the makefile and main.cpp to eliminate named modules.

Follow C++ practices

  • Class fields should be private, use getters and setters where necessary.
  • Methods not needed to be called externally should be private.
  • Try not to use multiple inheritance.
  • Make destructors virtual.
  • Use std::STL where appropriate.
  • Use C++11 where appropriate.

Code formatting

Must compile cleanly

Your contributed code must compile cleanly with no warnings.

Test your code!!

This is really important.

If you make changes to existing code make sure it is thoroughly tested before issuing a pull request.

If it is a new feature it should be tested as well as possible, or request testing by pointing to your branch in your repo.

Keep pull requests small

Only one feature or bug fix or refactor should go into a specific pull request, it is very hard to scan a huge pull request and to test it, so keep the changes small and testable.

Please follow git flow practices where possible.

Branches should be called feature/feature-name or fix/fix-name, and only contain changes relevant to that feature or fix.

If other changes need to be made that are not directly related to this change but this change relies on them then it should be noted in the pull request that it depends on pull request #xxxx.

Do not use blocking wait loops

There must not be any blocking wait loops in the code, there is no RTOS here so a blocking wait loop stops everything from running, and if long enough will trigger the watchdog timer.

In the case of needing to wait for several milliseconds you can use the safe_delay_ms or safe_delay_us function call, which will call on_idle() to keep things going.

However, note this is not an accurate delay. For accurate timing, an interrupt timer must be used.