WebRTC coding style guide

General advice

Some older parts of the code violate the style guide in various ways.

  • If making small changes to such code, follow the style guide when it’s reasonable to do so, but in matters of formatting etc., it is often better to be consistent with the surrounding code.
  • If making large changes to such code, consider first cleaning it up in a separate CL.

C++

WebRTC follows the Chromium and Google C++ style guides. In cases where they conflict, the Chromium style guide trumps the Google style guide, and the rules in this file trump them both.

C++ version

WebRTC is written in C++11, but with some restrictions:

  • We only allow the subset of C++11 (language and library) in the “allowed” section of this Chromium page.
  • We only allow the subset of C++11 that is also valid C++14; otherwise, users would not be able to compile WebRTC in C++14 mode.

Abseil

You may use a subset of the utilities provided by the Abseil library when writing WebRTC C++ code. Details.

.h and .cc files come in pairs

.h and .cc files should come in pairs, with the same name (except for the file type suffix), in the same directory, in the same build target.

  • If a declaration in path/to/foo.h has a definition in some .cc file, it should be in path/to/foo.cc.
  • If a definition in path/to/foo.cc file has a declaration in some .h file, it should be in path/to/foo.h.
  • Omit the .cc file if it would have been empty, but still list the .h file in a build target.
  • Omit the .h file if it would have been empty. (This can happen with unit test .cc files, and with .cc files that define main.)

This makes the source code easier to navigate and organize, and precludes some questionable build system practices such as having build targets that don’t pull in definitions for everything they declare.

Examples and exceptions.

ArrayView

When passing an array of values to a function, use rtc::ArrayView whenever possible—that is, whenever you’re not passing ownership of the array, and don’t allow the callee to change the array size.

For example,

instead ofuse
const std::vector<T>&ArrayView<const T>
const T* ptr, size_t num_elementsArrayView<const T>
T* ptr, size_t num_elementsArrayView<T>

See the source for more detailed docs.

absl::optional<T> as function argument

absl::optional<T> is generally a good choice when you want to pass a possibly missing T to a function—provided of course that T is a type that it makes sense to pass by value.

However, when you want to avoid pass-by-value, generally do not pass const absl::optional<T>&; use const T* instead. const absl::optional<T>& forces the caller to store the T in an absl::optional<T>; const T*, on the other hand, makes no assumptions about how the T is stored.

sigslot

sigslot is a lightweight library that adds a signal/slot language construct to C++, making it easy to implement the observer pattern with minimal boilerplate code.

When adding a signal to a pure interface, prefer to add a pure virtual method that returns a reference to a signal:

sigslot::signal<int>& SignalFoo() = 0;

As opposed to making it a public member variable, as a lot of legacy code does:

sigslot::signal<int> SignalFoo;

The virtual method approach has the advantage that it keeps the interface stateless, and gives the subclass more flexibility in how it implements the signal. It may:

  • Have its own signal as a member variable.

  • Use a sigslot::repeater, to repeat a signal of another object:

    sigslot::repeater<int> foo_;
    /* ... */
    foo_.repeat(bar_.SignalFoo());
    
  • Just return another object's signal directly, if the other object's lifetime is the same as its own.

    sigslot::signal<int>& SignalFoo() { return bar_.SignalFoo(); }
    

std::bind

Don’t use std::bind—there are pitfalls, and lambdas are almost as succinct and already familiar to modern C++ programmers.

std::function

std::function is allowed, but remember that it’s not the right tool for every occasion. Prefer to use interfaces when that makes sense, and consider rtc::FunctionView for cases where the callee will not save the function object.

Forward declarations

WebRTC follows the Google C++ style guide with respect to forward declarations. In summary: avoid using forward declarations where possible; just #include the headers you need.

C

There’s a substantial chunk of legacy C code in WebRTC, and a lot of it is old enough that it violates the parts of the C++ style guide that also applies to C (naming etc.) for the simple reason that it pre-dates the use of the current C++ style guide for this code base.

  • If making small changes to C code, mimic the style of the surrounding code.
  • If making large changes to C code, consider converting the whole thing to C++ first.

Java

WebRTC follows the Google Java style guide.

Objective-C and Objective-C++

WebRTC follows the Chromium Objective-C and Objective-C++ style guide.

Python

WebRTC follows Chromium’s Python style.

Build files

The WebRTC build files are written in GN, and we follow the Chromium GN style guide. Additionally, there are some WebRTC-specific rules below; in case of conflict, they trump the Chromium style guide.

WebRTC-specific GN templates

Use the following GN templates to ensure that all our targets are built with the same configuration:

instead ofuse
executablertc_executable
shared_libraryrtc_shared_library
source_setrtc_source_set
static_libraryrtc_static_library
testrtc_test

Target visibility and the native API

The WebRTC-specific GN templates declare build targets whose default visibility allows all other targets in the WebRTC tree (and no targets outside the tree) to depend on them.

Prefer to restrict the visibility if possible:

  • If a target is used by only one or a tiny number of other targets, prefer to list them explicitly: visibility = [ ":foo", ":bar" ]
  • If a target is used only by targets in the same BUILD.gn file: visibility = [ ":*" ].

Setting visibility = [ "*" ] means that targets outside the WebRTC tree can depend on this target; use this only for build targets whose headers are part of the native API.

Conditional compilation with the C preprocessor

Avoid using the C preprocessor to conditionally enable or disable pieces of code. But if you can’t avoid it, introduce a GN variable, and then set a preprocessor constant to either 0 or 1 in the build targets that need it:

if (apm_debug_dump) {
  defines = [ "WEBRTC_APM_DEBUG_DUMP=1" ]
} else {
  defines = [ "WEBRTC_APM_DEBUG_DUMP=0" ]
}

In the C, C++, or Objective-C files, use #if when testing the flag, not #ifdef or #if defined():

#if WEBRTC_APM_DEBUG_DUMP
// One way.
#else
// Or another.
#endif

When combined with the -Wundef compiler option, this produces compile time warnings if preprocessor symbols are misspelled, or used without corresponding build rules to set them.