Indent Style

From Aurora Free Open Source Software Wiki
Jump to: navigation, search

This document explains the codestyle that the Aurora Free Open Source Software uses. We use this rules to keep our code spaghetti safe and with a better look.

Aurora Codestyle Specification[edit]

This specification is based on K&R variants: 1TBS, Stroustrup, Linux kernel and BSD KNF codestyles.


In Aurora code files, we have a very well defined anatomy for comment blocks. You have three types of comments: the initial comments, documentation comments and code comments. All of them need to be written in English and well explained for easier development of other programmers that want to contribute to the project.

Initial Comments[edit]

Initial comments have general information about the file and should be present on ALL source code files. We have a boilerplate for C-like languages and languages that have # as 'keyword' for comment blocks. Depending on the license of the repository, you should use GPL-3.0 and LGPL-3.0 boilerplates. You can find them here.

Documentation Comments[edit]

To compile our documentation we use Doxygen, which uses JavaDoc Style for C-like languages and Doxygen style for other languages as a comment standard for automatically code structure detection. When you write a code file you should use this structure on header files:

1 /** Example comment
2  * Another comment
3  *
4  */

or this, in case of a non C-like language:

1 ## Example comment
2 # Another comment
3 #

You should use special commands so Doxygen knows what you want to do with the comment. Here are some useful and most used commands:

  • @file
  • @param
  • @return
  • @see
  • @author

You can find the full list here.

For better codestyle we activated the JAVADOC_AUTOBRIEF, which means that you can automatically initialize a brief description ending with a first dot followed by a space or a new line. For example:

1 /** Brief description which ends at this dot. Details follow
2  * here.
3  */

Note: Because of the aesthetics, use the brief in the first line (follow the example) whenever possible.

Each file must have a brief description, a @file tag, which has a partial path of the file and a @since tag to determinate when this file was created (framework version). It can also have the general author and contributers of the file, using @author or @authors.

@file tag must have this syntax: module-name/src/File.cpp or AuroraFW/Module/Header.h.

Each function must have an initial comment which explains:

  • What its purpose or use is.
  • What arguments it requires or accepts and what their use is.
  • What value(s) the function returns.
  • The version number of the framework since the first implementation.

It can also have the creator of the function, but its not mandatory.

In these comments you should use a formal and clear language, because the purpose of those comments are to automatically create the documentation reference of the framework.

Here is a practical example of a documentation comment:

1 /** Constructs a vector with the given coordinates.
2  * @param x The x value for the x coordinate.
3  * @param y The y value for the y coordinate.
4  * @see Vector2D( )
5  * @see Vector2D(float )
6  */

You can see the result here.

You can also use single comments for documentation: ///.

Code Comments[edit]

These types of comments can be used on headers and source files. The purpose of these comments is to help people understand an instruction or even a block of code. We use the common syntax of a comment: // and:

1 /*
2 block comments
3 */

Here you just need to be clear on what you want to explain to the programmer, so no formal language is needed and, for better communication, use easy words.


To name our code we use typical methods adopted in programming to know what's the type of the instructions used. We use a different name idention for types, functions and objects/variables.


For types such as a class, struct, typedef, enum, namespace, etc, you should use PascalCase, which means that the first letter of each concatenated word is capitalized. Here are some examples:

  • BackColor
  • TimeUtc
  • Timer


For functions you should use camelCase, which means that the first word is in lowercase and the rest of the words start with a capital letter. Here are some examples:

  • getName()
  • setName()
  • isNull()

Objects and Variables[edit]

For objects or variables we write them in lowercase. Specifically, for private members use an _ at the begin.

Note: Words that are already assigned by the language (keywords) should have an _ at the end.


If the first letter is uppercase then the whole acronym should have uppercase letters. Else if the first letter is lowercase then the whole acronym should have lowercase letters. Here are some examples:

  • ASCIIArt
  • asciiArt


Line structure is important for programmers. If lines are separated and with a length limit, code readability is improved and the development workflow is way better. It's important that blank lines exist and may be added to separate different blocks of code. It doesn't affect the compiler in any way, so there's no excuse for not using them.


When talking about soft limit, the lines should not pass 80 characters. For hard limit, lines must not pass 120 characters.


For line ending we use Unix LF (linefeed). If you are developing on Windows, you should use a linefeed compatible editor.

File encoding[edit]

You must use 8-bit unicode, UTF-8.


If it's a function, don't open braces in the same line as the declarations, else please open it in the same line. To close braces you must always do it in a new line, unless its has no body.

Single and multiple statements[edit]

Do not unnecessarily use braces where a single statement will do.

1 if (condition)
2 	action();


1 if (condition)
2 	do_this();
3 else
4 	do_that();

This does not apply if only one branch of a conditional statement is a single statement; in the latter case use braces in both branches:

1 if (condition) {
2 	do_this();
3 	do_that();
4 } else {
5 	otherwise();
6 }

Tab idention[edit]

Use tab instead of spaces for tab idention, if supported in the language, and configure your editor for 4 spaces in a single tab. Then, for alignment, use spaces. This helps to reduce significantly the project size.


Use a space after these keywords: if, switch, case, for, do, while, but not with sizeof, typeof, alignof, or __attribute__. Also don't add spaces around (inside) parenthesized expressions.


When declaring pointer data or a function that returns a pointer type, the preferred use of * is adjacent to the type name and not adjacent to the data or function name. For example:

1 char* name;
2 unsigned int memory(char* ptr, char** retptr);
3 char* convert(string* s);


In header files, you should write functions without argument names, or if needed, use very short names, like s, t, str, n, i, etc. Always use the global header on ALL headers (AuroraFW/Global.h). Use aurora defined data types, like ArInt_t, ArInt8_t, etc.


Names of macros defining constants and labels in enums are capitalized. Enums are preferred when defining several related constants. CAPITALIZED macro names are appreciated but macros resembling functions may be named in lower case. Generally, inline functions are preferable to macros resembling functions.

Macros with multiple statements should be enclosed in a do - while block:

1 #define macrofun(a, b, c)      \
2 	do {                   \
3 		if (a == 5)    \
4 		do_this(b, c); \
5 	} while (0)

Global macros must be defined on headers and shouldn't be used on source files (only if strictly needed).