Skip to content

Latest commit

 

History

History
161 lines (149 loc) · 5.83 KB

STYLEGUIDE.md

File metadata and controls

161 lines (149 loc) · 5.83 KB

Styleguide

Parentheses

Parentheses after function calls are to be placed right after function call without any spaces

functionName(parameters);

Parentheses after control statements (if, while, for, switch) are to be separated from these statements by a single space

if (something) ...;
while (something) ...;
for (something) ...;
switch (something) ...;

Code blocks

Two styles are used for blocks of code (enclosed by the curly braces {}). For if-else clauses and loops with only one call within, single-line style is used, (for the if-else clauses) usually with the matching calls indented in such way, that they start at the same column

if (something)           call(...);
else if (something_else) call(...);
else                     call(...);

In this case, the parameters should, if possible, be also aligned by the column. This isn't necessarily the case for loops or if-else clauses with different calls

if (something) call(...);
else if (something_else) otherCall(...);
else yetAnotherCall(...);
while (something) call(...);
for (something) call(...);

When a single line statement isn't sufficient, or a function/struct needs to be defined, the Allman style is used

if (something)
{
    call1(...);
    call2(...);
}
else
{
    call3(...);
    call4(...);
}
while (something)
{
    call1(...);
    call2(...);
}
type functionName(type parameter1, type parameter2)
{
    // do something
}

placing the curly brace on a new line.

These styles can be mixed and matched

if (something)
{
    call1(...);
    call2(...);
}
else if (something else) call(...);
else                     call(...);

Comments

Simple comments

Comments describing a block of calls should be placed a line above that "block", separated by an empty line with the previous. Comments, describing effects of specific calls should be placed besides the corresponding call and should be aligned in the same column. There should always be (at least) a single space between // and the actual comment.

something(...);
somethingElse(...);

// description of the call block
something(...);                    // description of the effect of `something`
somethingElse(...);                // description of the effect of `somethingElse`
whatever(...);

Multi-line comments

When describing more complicated behaviour, multi-line comments can be used. A multi-line comment should be structured as such

/*
    general description of some action
    can be written in sentences, if too long,
    should be divided into multiple lines
        - lists
        - can be
        - utilized too
*/

The /* */ marks should always be placed at the current indentation level. 1 soft-tab should be used for the description, 2 soft-tabs should be used for lists.

Commenting out function calls

When commenting out function calls, two slashes with one space (// ) must be used. This is because many text editors support Ctrl + / shortcut for quick commenting and uncommenting selected calls. This feature sadly doesn't work with multi-line comments, thus single line comments should always be used in this case.

// inactive block
// something(...);
// somethingElse(...);

// description of the call block
something(...);                    // description of the effect of `something`
somethingElse(...);                // description of the effect of `somethingElse`
whatever(...);

Documenting functions

When writing a reference comment for a defined function, the following template should be used

/*
    function description (what it does)
        parameter1 - description of parameter1
        parameter2 - description of parameter2
        ...
        @return    - description of the returned value (can be left out if the return type is void)
*/

Suppose a function for getting bar from foo is defined (foo being its only parameter), then the definition in some .h file with the reference comment would be

/*
    gets bar from foo and returns it
        f       - foo object
        @return - bar object
*/
bar fooGetBar(foo f);

Naming

Functions

To name functions, camelCase is used. The first character of the name is to be lower cased. Any other word, appended to the name is to have its first character capitalized. When naming functions, tightly related to specific types, the name of the type should be the first word to appear in the function name.

Suppose a type foo is declared, then the constructor for foo is to be declared as

foo fooNew(...);

and a function doing action get bar (returning type bar) on foo should be declared as

bar fooGetBar(foo f, ...);

taking some foo (or a pointer to some foo) as its first parameter.

Structs and types

Ideally, the struct and type names be single words. If such is not possibl, the struct name should always be expressed in a PascalCase, having the first character in its name capitalized. The corresponding type name should be then expressed as a camelCase of the struct name (to comply with function naming scheme).

Suppose a foo bar structure is to be defined. With the given conventions, it would be defined as such

typedef struct FooBar      // struct name
{
    // fields
}
fooBar;                    // type name

Variables and fields

Variables and fields are perhaps the most disorganized category in this code. There are only three rules (but mostly just recommendations), really,

  • variables must not be named using camelCase, as one could accidentally name a variable identically to an already defined function,
  • when using single letter names, the meaning of the variable must be either painfully obvious from the context or must be explained using a comment,
  • for longer variable names with multiple words, snake_case should be used (e.g. very_long_variable_name_please_try_to_avoid_these).