GOLDEN RULE: Follow the style of the existing code when you make changes.
- Use tabs for leading indentation:
- tab stops are every 4 characters (only relevant for line length).
- one indentation level -> exactly one byte (i.e. a tab character) in the source file.
- Line widths:
- Lines should be at most 99 characters wide to make diff views readable and reduce merge conflicts.
- Lines of comments should be formatted according to ease of viewing, but simplicity is to be preferred over beauty.
- Single-statement blocks should not have braces, unless required for clarity.
- Never place condition bodies on same line as condition.
- Space between keyword and opening parenthesis, but not following opening parenthesis or before final parenthesis.
- No spaces for unary operators,
->
or.
. - No space before
:
but one after it, except in the ternary operator: one on both sides. - Add spaces around all other operators.
- Braces, when used, always have their own lines and are at same indentation level as "parent" scope.
- If lines are broken, a list of elements enclosed with parentheses (of any kind) and separated by a separator (of any kind) are formatted such that there is exactly one element per line, followed by the separator, the opening parenthesis is on the first line, followed by a line break and the closing parenthesis is on a line of its own unindented). See example below.
Yes:
if (a == b[i])
printf("Hello\n"); // NOTE spaces used instead of tab here for clarity - first byte should be '\t'.
foo->bar(
someLongVariableName,
anotherLongVariableName,
anotherLongVariableName,
anotherLongVariableName,
anotherLongVariableName
);
cout <<
"some very long string that contains completely irrelevant " <<
"text that talks about this and that and contains the words " <<
"\"lorem\" and \"ipsum\"" <<
endl;
No:
if( a==b[ i ] ) { printf ("Hello\n"); }
foo->bar(someLongVariableName,
anotherLongVariableName,
anotherLongVariableName,
anotherLongVariableName,
anotherLongVariableName);
cout << "some very long string that contains completely irrelevant text that talks about this and that and contains the words \"lorem\" and \"ipsum\"" << endl;
To set indentation and tab width settings uniformly, the repository contains an EditorConfig .editorconfig
file, which describes some of the styles used and which is recognized by many IDE's and editors.
- No
using namespace
declarations in header files. - Use
using namespace std;
in cpp files, but avoid importing namespaces from boost and others. - All symbols should be declared in a namespace except for final applications.
- Use anonymous namespaces for helpers whose scope is a cpp file only.
- Preprocessor symbols should be prefixed with the namespace in all-caps and an underscore.
- Do not use
std::
qualifier in cpp files (see 2.), except forstd::move
, which will otherwise cause thecheck_style
step to fail.
Only in the header:
#include <cassert>
namespace myNamespace
{
std::tuple<float, float> meanAndSigma(std::vector<float> const& _v);
}
Only in the cpp file:
#include <cassert>
using namespace std;
tuple<float, float> myNamespace::meanAndSigma(vector<float> const& _v)
{
// ...
}
- File comment is always at top, and includes:
- Copyright
- License (e.g. see COPYING)
- Never use
#ifdef
/#define
/#endif
file guards. Prefer#pragma
once as first line below file comment. - Prefer static constexpr variables to value macros.
- Prefer inline constexpr functions to function macros.
- Split complex macro on multiple lines with
\
.
GOLDEN RULE: Preprocessor: ALL_CAPS
; C++: camelCase
.
- Use camelCase for splitting words in names, except where obviously extending STL/boost functionality in which case follow those naming conventions.
- The following entities' first alpha is upper case:
- Type names
- Template parameters
- Enum members
- static const variables that form an external API.
- All preprocessor symbols (macros, macro arguments) in full uppercase with underscore word separation.
All other entities' first alpha is lower case.
- Leading underscore "_" to parameter names:
- Exception: "o_parameterName" when it is used exclusively for output. See 6(f).
- Exception: "io_parameterName" when it is used for both input and output. See 6(f).
- Leading "g_" to global (non-const) variables.
- Leading "s_" to static (non-const, non-global) variables.
Use solAssert
and solUnimplementedAssert
generously to check assumptions that span across different parts of the code base, for example before dereferencing a pointer.
- {Typename} + {qualifiers} + {name}.
- Only one per line.
- Associate */& with type, not variable (at ends with parser, but more readable, and safe if in conjunction with (b)).
- Favour declarations close to use; don't habitually declare at top of scope ala C.
- Pass non-trivial parameters as const reference, unless the data is to be copied into the function, then either pass by const reference or by value and use std::move.
- If a function returns multiple values, use std::tuple (std::pair acceptable) or better introduce a struct type. Do not use */& arguments.
- Use parameters of pointer type only if
nullptr
is a valid argument, use references otherwise. Often,std::optional
is better suited than a raw pointer. - Never use a macro where adequate non-preprocessor C++ can be written.
- Only use
auto
if the type is very long and rather irrelevant. - Do not pass bools: prefer enumerations instead.
- Prefer enum class to straight enum.
- Always initialize POD variables, even if their value is overwritten later.
Yes:
enum class Accuracy
{
Approximate,
Exact
};
struct MeanSigma
{
float mean = 0.0f;
float standardDeviation = 1.0f;
};
double const d = 0;
int i = 0;
int j = 0;
char* s = nullptr;
MeanAndSigma ms meanAndSigma(std::vector<float> const& _v, Accuracy _a);
Derived* x = dynamic_cast<Derived*>(base);
for (auto i = x->begin(); i != x->end(); ++i) {}
No:
const double d = 0;
int i, j;
char *s;
float meanAndSigma(std::vector<float> _v, float* _sigma, bool _approximate);
Derived* x(dynamic_cast<Derived*>(base));
for (map<ComplexTypeOne, ComplexTypeTwo>::iterator i = l.begin(); i != l.end(); ++l) {}
- Structs to be used when all members public and no virtual functions:
- In this case, members should be named naturally and not prefixed with
m_
.
- In this case, members should be named naturally and not prefixed with
- Classes to be used in all other circumstances.
- One member per line only.
- Private, non-static, non-const fields prefixed with
m_
. - Avoid public fields, except in structs.
- Use override, final and const as much as possible.
- No implementations with the class declaration, except:
- template or force-inline method (though prefer implementation at bottom of header file).
- one-line implementation (in which case include it in same line as declaration).
- For a property
foo
- Member:
m_foo
- Getter:
foo()
[ also: for booleans,isFoo()
] - Setter:
setFoo()
- Member:
- Avoid unpronounceable names.
- Names should be shortened only if they are extremely common, but shortening should be generally avoided
- Avoid prefixes of initials (e.g. do not use
IMyInterface
,CMyImplementation
) - Find short, memorable & (at least semi-) descriptive names for commonly used classes or name-fragments:
- A dictionary and thesaurus are your friends;
- Spell correctly;
- Think carefully about the class's purpose;
- Imagine it as an isolated component to try to decontextualise it when considering its name;
- Don't be trapped into naming it (purely) in terms of its implementation.
- Prefer
using
totypedef
. e.g.using ints = std::vector<int>;
rather than typedefstd::vector<int> ints;
- Generally avoid shortening a standard form that already includes all important information:
- e.g. stick to
shared_ptr<X>
rather than shortening toptr<X>
.
- e.g. stick to
- Where there are exceptions to this (due to excessive use and clear meaning), note the change prominently and use it consistently:
- e.g.
using Guard = std::lock_guard<std::mutex>;
///< Guard is used throughout the codebase since it is clear in meaning and used commonly.
- e.g.
- In general expressions should be roughly as important/semantically meaningful as the space they occupy.
- Avoid introducing aliases for types unless they are very complicated. Consider the number of items a brain can keep track of at the same time.
- Comments should be doxygen-compilable, using @notation rather than \notation.
- Document the interface, not the implementation:
- Documentation should be able to remain completely unchanged, even if the method is reimplemented;
- Comment in terms of the method properties and intended alteration to class state (or what aspects of the state it reports);
- Be careful to scrutinise documentation that extends only to intended purpose and usage;
- Reject documentation that is simply an English transaction of the implementation.
- Avoid in-code comments. Instead, try to extract blocks of functionality into functions. This often already eliminates the need for an in-code comment.
- Includes should go in increasing order of generality (
libsolidity
->libevmasm
->libsolutil
->boost
->STL
). - The corresponding
.h
file should be the first include in the respective.cpp
file. - Insert empty lines between blocks of include files.
Example:
#include <libsolidity/codegen/ExpressionCompiler.h>
#include <libsolidity/ast/AST.h>
#include <libsolidity/codegen/CompilerContext.h>
#include <libsolidity/codegen/CompilerUtils.h>
#include <libsolidity/codegen/LValue.h>
#include <libevmasm/GasMeter.h>
#include <libsolutil/Common.h>
#include <libsolutil/SHA3.h>
#include <boost/range/adaptor/reversed.hpp>
#include <boost/algorithm/string/replace.hpp>
#include <utility>
#include <numeric>
See this issue for the reason: this makes it easier to find missing includes in header files.
-
Herb Sutter and Bjarne Stroustrup:
-
Herb Sutter and Andrei Alexandrescu:
- "C++ Coding Standards: 101 Rules, Guidelines, and Best Practices"
-
Scott Meyers:
- "Effective C++: 55 Specific Ways to Improve Your Programs and Designs (3rd Edition)"
- "More Effective C++: 35 New Ways to Improve Your Programs and Designs"
- "Effective Modern C++: 42 Specific Ways to Improve Your Use of C++11 and C++14"