Skip to content

Coding Standards

Shintaro Iwasaki edited this page Mar 9, 2020 · 4 revisions

General Principles

Coding standards are intended to make this project more successful and enjoyable. We now use our custom coding style in favor of clang-format (https://clang.llvm.org/docs/ClangFormat.html). This style is based on K&R style with four spaces for an indent.

Before Creating a PR

The Argobots project uses a style checker. Before you create a PR, please pass the following script:

maint/code-cleanup.sh YOURCODE

If you want to perform this clean-up script through all your code files in the current directory, type as follows:

maint/code-cleanup.sh --all --recursive

You will find a test failure if your PR does not comply with this style checker.

Basic Source File Structure

All source files follow similar structure; the file maint/template.c gives an example.

#include "abti.h"

int ABTI_template(ABTI_thread *p_thread, void *p_arg)
{
    int abt_errno = ABT_SUCCESS;

    if (p_thread == NULL) {
        HANDLE_ERROR("NULL THREAD");
        abt_errno = ABT_ERR_INV_THREAD;
        goto fn_fail;
    }

    /* Implementation */
    /* ... */

fn_exit:
    return abt_errno;

fn_fail:
    goto fn_exit;
}

Style header

This sets the style for the file and sets a common indentation amount for C and C++ programs. In addition, the style header is important for C++ header files so that the style checker will know that the file is C++ instead of C.

/* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil ; -*- */

Copyright

COPYRIGHT file exists in top-level directory. If the default copyright is applied to the source file, the following can be used:

/*
 * See COPYRIGHT in top-level directory.
 */

Header files

Header files should have their contents wrapped within an ifndef of the form

#ifndef FILENAME_H_INCLUDED
#define FILENAME_H_INCLUDED
...
#endif /* FILENAME_H_INCLUDED */

where FILENAME is the name of the file, in upper case.

Column width

Lines within files should be limited to 80 columns.

Comments

Use /* */ and avoid // for writing comments. It allows compilers not supporting C99 to compile the source file with no problem on comments. If you want to write comments for doxygen documentation, refer to Code Documentation.

Code cleanup script

The script maint/code-cleanup.sh can be used to cleanup whitespace, comments, and line-wrapping in existing source files. It is a good idea to run it on any newly created files in the git repo.

Naming Convention

Naming is important to distinguish internal implementation from public interface. Argobots uses prefix rules to differentiate codes according to their intention, i.e., public or internal. All characters in names except the argobots prefix should be lower-case. For example,

int ABT_thread_create()

After ABT_, characters are all in lower case.

Type and routine names

Names for types and routines should be chosen in a way that (1) is descriptive of their function and (2) is clearly distinct from any name that may be used by another runtime library or user code. To achieve this, we have chosen to reserve a few prefixes for the use of the Argobots implementation:

Prefix Description
ABT_ This is used for public interface that is exposed to users. All public data types and APIs should start with this prefix.
ABTI_ This is used for internal data types and routines to implement the Argobots public APIs. This does not include device-specific types and routines.
ABTD_ This is used for device-specific data types and routines to implement ABT_ and ABTI_ routines. It plays a role of an interface to hide internal device-specific implementation.
ABTDI_ This is used for internal implementation for device-specific types and routines. It implements ABTD_ types and routines according to the target architecture. The configure script should correctly choose an appropriate implementation for the target architecture.
ABTU_ This is used for utility routines, which support general functions not only limited in implementation of Argobots. Therefore, utility routines should be able to be compiled independently.

Macro and enum names

Just like type and routine names, it is important to pick macro and enum names to avoid possible conflicts with others defined in system header files.

Prefix/Suffix Description
ABT_ Used for public interface.
ABTI_, ABTD_, ABTDI_, ABTU_ Used for internal implementation. Each prefix has same meaning as one in type and routine names.
DEBUG_ Used for debugging.
_H_INCLUDED This suffix should be used as the test for including a header (.h) file.

Every macro and enum name starting with above prefixes should be all uppercase.

Variables

Some prefixes are used to distinguish global variables and pointers. For regular variables, currently there is no strict rule for naming them.

Prefix Description
g_ Used for global variables
gp_ Used for global pointers
l_ Used for ES local variables, which are shared only in an ES
lp_ Used for ES local pointers
p_ Used for pointer variables
f_ Used for function pointers

Error Handling

TBD

Memory Allocation

To manage dynamically-allocated memory in Argobots, we use a special set of memory allocation macros that are treated as utility routines. The macro names and their corresponding functions are the following:

Macro Name Corresponding Function
ABTU_malloc malloc
ABTU_calloc calloc
ABTU_free free
ABTU_realloc realloc

The definitions are defined in src/abtu.h. We are planning to exploit these macros to test and diagnose memory corruption and leak problems when they occur.