1. Coding conventions

This code is used by many people and therefore you should keep some things in mind when you submit source code:

• Keep things simple
• Keep the source clean
• Always know what you are doing
• Tell what you are doing
• Remember that you write code once but that it is read many times by many people

AROS uses some of the comments in the source to generate the documentation. Therefore it's neccessary to keep a certain format so the tools can find their information. Other comments are ignored but they should explain what you thought when you wrote the code. If you really can't think of an explanation, then don't write the code a second time like this:

What we think of is this:

Every function in AROS must have a full ANSI C prototype. Prototypes should be collected in in one header per file if it is needed by only a few files (no need to recompile the whole project if you change a function which used only once), in one header per directory if it's a commonly used function in that directory or in one header per logical group (ie. one header for all functions in a library).

The function header (ie. the comment before the function) must be of a special format because the AutoDocs are generated from it. Here is an example for it (from AROS/exec/addhead.c):

As you can see, comments are used to merge the function prototype and the header into one.

NAME

This field contains all neccessary prototypes to use the function from the user point of view and the name of the function in a AROS_LH*() macro (Library Header). These macros are used to make the same code work on different kind of hardwares. The name of the macro depends on the amount of parameters and whether the function needs the library base. AddHead() does not and therefore an "I" is appended to the macros name. If it need the library base (like AddTask()), then the "I" is omitted.

If the function is not part of a shared library and it's arguments must be passed in certain registers (eg. callback hooks), you must use AROS_UFH*() macros (User Function Header) instead of AROS_LH*(). Append the number of arguments to this macro. Since it has never a base, the field LOCATION must be omitted and it's not neccessary to append the "I" to the macros name. An example for a callback hook foo() would be:

AROS_UFH3(ULONG, foo,
    AROS_UFHA(struct Hook, hook,  A0),
    AROS_UFHA(APTR,        obj,   A2),
    AROS_UFHA(APTR,        param, A1)
)

(note that the registers need not have a particular order).

If the function is not part of a shared library and it's arguments need not be in specific registers, you need no AROS_*H*() macros:

/*****************************************************************************

    NAME */
#include <header.h>

        int foo (

/*  SYNOPSIS */
        int a,
        int b)

/*  FUNCTION
        blahblahblah.
        ...

*****************************************************************************/

SYNOPSIS

This field contains all arguments of the function one by one in AROS_LHA() macros (Library Header Argument). This macro makes sure the respective argument is put in the right CPU register when the function is called (if possible and neccessary). The first argument for the macro is the type of the parameter followed by the name of the parameter and the register the parameter is expected in. Valid names for registers are D0, D1, D2 upto D7 and A0 upto A6.

If the function is not part of a library but the arguments must be passed to it in registers, then use AROS_UFHA() macros (User Function Header Argument) which take the same parameters as the AROS_LHA() macros. Don't forget the closing parenthese for the AROS_UFC.

If the function is not part of a library and the arguments need not be passed in registers, no macros are neccessary.

LOCATION

This field is neccessary for shared libraries only. It contains the last four parameters for the AROS_LH*() macro which are the type of the library, the name of the variable, in which the function expects the library base, the offset of the function in the jumptable (the first vector has 1 and the first vector which may be used by a function is 5) and the name of the library.

FUNCTION

This field contains a description of the function.

INPUTS

This field contains a list of all parameters of the form "name - description" or "name, name, name - description". The description should tell what the parameter is and what values can be passed to it. There is no point in explaining the parameter twice in FUNCTION and here. If the function has no parameters, say "None." here.

RESULT

What the function passes back. This includes return values and values passed in arguments of the function. If the function may fail, you should explain what it returns on failure and why it might fail.

NOTES

Important things the user must know or take into account.

EXAMPLE

This field should contain a small or fully featured example. A good way to present an example is to write some code which tests the function, put it into #ifdef TEST somewhere in the file and put a "See below." here. If you need comments in the code, you have two ways for this. If you need only short one-line comments, use C++ style (|// comment). Everything from the //| to the end of the line is the comment. If you need more comment, then you can end the comment after the EXAMPLE and use #ifdef EXAMPLE to mask the example out:
    EXAMPLE */
#ifdef EXAMPLE
        struct List * list;
        struct Node * pred;

        /* Insert Node at top of the list */
        AddHead (list, node);
#endif
Don't use #ifdef EXAMPLE if you have a fully featured example (ie. one which can be compiled without errors).

BUGS

This field contains a list of known bugs.

SEE ALSO

This field contains a list of other functions and documents which might be of interest. This includes function which you need to initialize, create or destroy an object necessary for this function, functions which do similar and opposite things on the main object. For example, SetAttrs() should contain functions here which can create, destroy and manipulate BOOPSI objects but not taglists.

INTERNALS

This field should contain information for other developers which are irrelevant to the user, for example an explanation of the algorithm of the function or dependencies.

HISTORY

Since we use CVS, this field is obsolete. You can delete it, if you encounter it.

Here is an example of how to format AROS code:

Looks ugly, eh ? :-) Ok, here are the rules:

• If several lines contain similar code, put similar things below each other (see a and b);
• Put spaces between operands and operators
• Put braces {}, brackets [] and parenthese () below each other (d) if there is much code between. Brackets and parenthese may be in one line if the code between is small (c)
• Indent by 4 Spaces. Two indent levels may be abbreviated by one tab.

  The reasons for this are: 1. While some editors can use an arbitrary sizes for tabs, it's a bit complicated to tell another editor which tab size was used by the one used to write the code. 2. Most code in AROS was written this way and your code should look like the rest. 3. You can print this code on any printer without special tools to "fix" the tabs. 4. Most editors have smart tabs which do exactly this. If your editor doesn't, write a bug report.

• If you have a function with many arguments (d, e) you should put the parenthese in lines of their own and each argument in one line (d) or put the first argument behind the opening parenthese (e) and each following argument in a line of its own with the comma in front. The closing parenthese is in a line of its own and aligned with the beginning of the expression (ie. the a and not the opening parenthese or the printf()).
• use a single blank line to separate logical blocks. Large comments should have a blank line before and after them, small comments should be put before the code they explain with only one blank line before them.

Code in AROS modules should be written in a way that makes it suitable for embedding into a ROM, FlashRAM or other kinds read-only memory. The following coding style rules are meant to make it possible. Of course they apply to all Kickstart modules and to code that may be made resident, shared or linked to other modules.

• ROM modules must have no .data and .bss sections. Basically, we need to get rid of all non-const global data. The real Amiga Kickstart proves that it's both possible and easy to achieve this. If you encounter an external variable (static or not) that is modified by the code, try to get rid of it or move it into the base of the library/device (or in the device node of your handler or in the userdata of your class).
• The above applies to library bases as well. If you are writing a library, put the bases of other libraries into your own library base structure. Boopsi classes can store library bases in their class private data.
• Try to set the static and const attributes to all your global data. You can also use the CONST_STRPTR and CONST_APTR types defined in <exec/types.h>. Using static const allows the compiler to move data into the .text (AKA code) segment. If you need to pass these globals to another function, try to chenge its prototype to use const too. Note that, as of OS 3.5, Olaf Barthel has finally switched to using const in <clib/#?_protos.h> headers.
• NEVER EVER touch buffers passed in by the user as an "input" parameter. The concept of input parameters is often implicit in the function description. For instance, the filename passed to Open() is clearly an input variable and Open() must not mess with it, even if it is going to fix it back later. Keep in mind that the buffer might be in read-only memory or shared among several instances of a resident or multithreaded program.
• Try to avoid host-OS calls such as malloc() and free() if you can do with AllocMem() and FreeMem(). This is because the pointer checking debug macros rely on finding the pointer within the Exec memory blocks with TypeOfMem().