riscv-openocd/doc/manual/style.txt

388 lines
16 KiB
Plaintext

/** @page styleguide Style Guides
The goals for each of these guides are:
- to produce correct code that appears clean, consistent, and readable,
- to allow developers to create patches that conform to a standard, and
- to eliminate these issues as points of future contention.
Some of these rules may be ignored in the spirit of these stated goals;
however, such exceptions should be fairly rare.
The following style guides describe a formatting, naming, and other
conventions that should be followed when writing or changing the OpenOCD
code:
- @subpage styletcl
- @subpage stylec
- @subpage styleperl
- @subpage styleautotools
In addition, the following style guides provide information for
providing documentation, either as part of the C code or stand-alone.
- @subpage styledoxygen
- @subpage styletexinfo
- @subpage stylelatex
Feedback would be welcome to improve the OpenOCD guidelines.
*/
/** @page styletcl TCL Style Guide
OpenOCD needs to expand its Jim/TCL Style Guide.
Many of the guidelines listed on the @ref stylec page should apply to
OpenOCD's Jim/TCL code as well.
*/
/** @page stylec C Style Guide
This page contains guidelines for writing new C source code for the
OpenOCD project.
@section styleformat Formatting Guide
- remove any trailing white space at the end of lines.
- use TAB characters for indentation; do NOT use spaces.
- displayed TAB width is 4 characters.
- use Unix line endings ('\\n'); do NOT use DOS endings ('\\r\\n')
- limit adjacent empty lines to at most two (2).
- remove any trailing empty lines at the end of source files
- do not "comment out" code from the tree; instead, one should either:
-# remove it entirely (Subversion can retrieve the old version), or
-# use an @c \#if/\#endif block.
Finally, try to avoid lines of code that are longer than than 72-80 columns:
- long lines frequently indicate other style problems:
- insufficient use of static functions, macros, or temporary variables
- poor flow-control structure; "inverted" logical tests
- a few lines may be wider than this limit (typically format strings), but:
- all C compilers will concatenate series of string constants.
- all long string constants should be split across multiple lines.
@section stylenames Naming Rules
- most identifiers must use lower-case letters (and digits) only.
- macros must use upper-case letters (and digits) only.
- OpenOCD identifiers should NEVER use @c MixedCaps.
- structure names must end with the '_s' suffix.
- typedef names must end with the '_t' suffix.
- use underline characters between consecutive words in identifiers
(e.g. @c more_than_one_word).
@section stylec99 C99 Rules
- inline functions
- @c // comments -- in new code, prefer these for single-line comments
- trailing comma allowed in enum declarations
- designated initializers (@{ .field = value @})
- variables declarations may be mixed with code
- new block scopes for selection and iteration statements
@section styletypes Type Guidelines
- use native types (@c int or @c unsigned) if the type is not important
- if size matters, use the types from \<stdint.h\> or \<inttypes.h\>:
- @c int8_t, @c int16_t, @c int32_t, or @c int64_t: signed types of specified size
- @c uint8_t, @c uint16_t, @c uint32_t, or @c uint64_t: unsigned types of specified size
- do @b NOT redefine @c uN types from "types.h"
@section stylefunc Functions
- static inline functions should be prefered over macros:
@code
/** do NOT define macro-like functions like this... */
#define CUBE(x) ((x) * (x) * (x))
/** instead, define the same expression using a C99 inline function */
static inline int cube(int x) { return x * x * x; }
@endcode
- Functions should be declared static unless required by other modules
- define static functions before first usage to avoid forward declarations.
- Functions should have no space between its name and its parameter list:
@code
int f(int x1, int x2)
{
...
int y = f(x1, x2 - x1);
...
}
@endcode
*/
/** @page styledoxygen Doxygen Style Guide
The following sections provide guidelines for OpenOCD developers
who wish to write Doxygen comments in the code or this manual.
For an introduction to Doxygen documentation,
see the @ref primerdoxygen.
@section styledoxyblocks Doxygen Block Selection
Several different types of Doxygen comments can be used; often,
one style will be the most appropriate for a specific context.
The following guidelines provide developers with heuristics for
selecting an appropriate form and writing consistent documentation
comments.
-# use @c /// to for one-line documentation of instances.
-# for documentation requiring multiple lines, use a "block" style:
@verbatim
/**
* @brief First sentence is short description. Remaining text becomes
* the full description block, where "empty" lines start new paragraphs.
*
* One can make text appear in @a italics, @b bold, @c monospace, or
* in blocks such as the one in which this example appears in the Style
* Guide. See the Doxygen Manual for the full list of commands.
*
* @param foo For a function, describe the parameters (e.g. @a foo).
* @returns The value(s) returned, or possible error conditions.
*/
@endverbatim
-# The block should start on the line following the opening @c /**.
-# The end of the block, \f$*/\f$, should also be on its own line.
-# Every line in the block should have a @c '*' in-line with its start:
- A leading space is required to align the @c '*' with the @c /** line.
- A single "empty" line should separate the function documentation
from the block of parameter and return value descriptions.
- Except to separate paragraphs of documentation, other extra
"empty" lines should be removed from the block.
-# Only single spaces should be used; do @b not add mid-line indentation.
-# If the total line length will be less than 72-80 columns, then
- The @c /**< form can be used on the same line.
- This style should be used sparingly; the best use is for fields:
@code int field; /**< field description */ @endcode
@section styledoxyall Doxygen Style Guide
The following guidelines apply to all Doxygen comment blocks:
-# Use the @c '\@cmd' form for all doxygen commands (do @b not use @c '\\cmd').
-# Use symbol names such that Doxygen automatically creates links:
-# @c function_name() can be used to reference functions
(e.g. flash_set_dirty()).
-# @c struct_name::member_name should be used to reference structure
fields in the documentation (e.g. @c flash_driver_s::name).
-# URLS get converted to markup automatically, without any extra effort.
-# new pages can be linked into the heirarchy by using the @c \@subpage
command somewhere the page(s) under which they should be linked:
-# use @c \@ref in other contexts to create links to pages and sections.
-# Use good Doxygen mark-up:
-# '\@a' (italics) should be used to reference parameters (e.g. <i>foo</i>).
-# '\@b' (bold) should be used to emphasizing <b>single</b> words.
-# '\@c' (monospace) should be used with <code>file names</code> and
<code>code symbols</code>, so they appear visually distinct from
surrounding text.
-# To mark-up multiple words, the HTML alternatives must be used.
-# Two spaces should be used when nesting lists; do @b not use '\\t' in lists.
-# Code examples provided in documentation must conform to the Style Guide.
@section styledoxytext Doxygen Text Inputs
In addition to the guidelines in the preceding sections, the following
additional style guidelines should be considered when writing
documentation as part of standalone text files:
-# Text files must contain Doxygen at least one comment block:
-# Documentation should begin in the first column (except for nested lists).
-# Do NOT use the @c '*' convention that must be used in the source code.
-# Each file should contain at least one @c \@page block.
-# Each new page should be listed as a \@subpage in the \@page block
of the page that should serve as its parent.
-# Large pages should be structure in parts using meaningful \@section
and \@subsection commands.
-# Include a @c \@file block at the end of each Doxygen @c .txt file to
document its contents:
- Doxygen creates such pages for files automatically, but no content
will appear on them for those that only contain manual pages.
- The \@file block should provide useful meta-documentation to assist
techincal writers; typically, a list of the pages that it contains.
- For example, the @ref styleguide exists in @c doc/manual/style.txt,
which contains a reference back to itself.
-# The \@file and \@page commands should begin on the same line as
the start of the Doxygen comment:
@verbatim
/** @page pagename Page Title
Documentation for the page.
*/
/** @file
This file contains the @ref pagename page.
*/
@endverbatim
For an example, the Doxygen source for this Style Guide can be found in
@c doc/manual/style.txt, alongside other parts of The Manual.
*/
/** @page styletexinfo Texinfo Style Guide
The User's Guide is there to provide two basic kinds of information. It
is a guide for how and why to use each feature or mechanism of OpenOCD.
It is also the reference manual for all commands and options involved
in using them, including interface, flash, target, and other drivers.
At this time, it is the only user-targetted documentation; everything
else is addressing OpenOCD developers.
There are two key audiences for the User's Guide, both developer based.
The primary audience is developers using OpenOCD as a tool in their
work, or who may be starting to use it that way. A secondary audience
includes developers who are supporting those users by packaging or
customizing it for their hardware, installing it as part of some software
distribution, or by evolving OpenOCD itself. There is some crossover
between those audiences. We encourage contributions from users as the
fundamental way to evolve and improve OpenOCD. In particular, creating
a board or target specific configuration file is something that many
users will end up doing at some point, and we like to see such files
become part of the mainline release.
General documentation rules to remember include:
- Be concise and clear. It's work to remove those extra words and
sentences, but such "noise" doesn't help readers.
- Make it easy to skim and browse. "Tell what you're going to say,
then say it". Help readers decide whether to dig in now, or
leave it for later.
- Make sure the chapters flow well. Presentations should not jump
around, and should move easily from overview down to details.
- Avoid using the passive voice.
- Address the reader to clarify roles ("your config file", "the board you
are debugging", etc.); "the user" (etc) is artificial.
- Use good English grammar and spelling. Remember also that English
will not be the first language for many readers. Avoid complex or
idiomatic usage that could create needless barriers.
- Use examples to highlight fundamental ideas and common idioms.
- Don't overuse list constructs. This is not a slide presentation;
prefer paragraphs.
When presenting features and mechanisms of OpenOCD:
- Explain key concepts before presenting commands using them.
- Tie examples to common developer tasks.
- When giving instructions, you can \@enumerate each step both
to clearly delineate the steps, and to highlight that this is
not explanatory text.
- When you provide "how to use it" advice or tutorials, keep it
in separate sections from the reference material.
- Good indexing is something of a black art. Use \@cindex for important
concepts, but don't overuse it. In particular, rely on the \@deffn
indexing, and use \@cindex primarily with significant blocks of text
such as \@subsection. The \@dfn of a key term may merit indexing.
- Use \@xref (and \@anchor) with care. Hardcopy versions, from the PDF,
must make sense without clickable links (which don't work all that well
with Texinfo in any case). If you find you're using many links,
read that as a symptom that the presentation may be disjointed and
confusing.
- Avoid font tricks like \@b, but use \@option, \@file, \@dfn, \@emph
and related mechanisms where appropriate.
For technical reference material:
- It's OK to start sections with explanations and end them with
detailed lists of the relevant commands.
- Use the \@deffn style declarations to define all commands and drivers.
These will automatically appear in the relevant index, and those
declarations help promote consistent presentation and style.
- It's a "Command" if it can be used interactively.
- Else it's a "Config Command" if it must be used before the
configuration stage completes.
- For a "Driver", list its name.
- Use BNF style regular expressions to define parameters:
brackets around zero-or-one choices, parentheses around
exactly-one choices.
- Use \@option, \@file, \@var and other mechanisms where appropriate.
- Say what output it displays, and what value it returns to callers.
- Explain clearly what the command does. Sometimes you will find
that it can't be explained clearly. That usually means the command
is poorly designed; replace it with something better, if you can.
- Be complete: document all commands, except as part of a strategy
to phase something in or out.
- Be correct: review the documentation against the code, and
vice versa.
- Alphabetize the \@defn declarations for all commands in each
section.
- Keep the per-command documentation focussed on exactly what that
command does, not motivation, advice, suggestions, or big examples.
When commands deserve such expanded text, it belongs elsewhere.
Solutions might be using a \@section explaining a cluster of related
commands, or acting as a mini-tutorial.
- Details for any given driver should be grouped together.
The User's Guide is the first place most users will start reading,
after they begin using OpenOCD. Make that investment of their time
be as productive as possible. Needing to look at OpenOCD source code,
to figure out how to use it is a bad sign, though it's OK to need to
look at the User's guide to figure out what a config script is doing.
*/
/** @page stylelatex LaTeX Style Guide
This page needs to provide style guidelines for using LaTeX, the
typesetting language used by The References for OpenOCD Hardware.
Likewise, the @ref primerlatex for using this guide needs to be completed.
*/
/** @page styleperl Perl Style Guide
This page provides some style guidelines for using Perl, a scripting
language used by several small tools in the tree:
-# Ensure all Perl scripts use the proper suffix (@c .pl for scripts, and
@c .pm for modules)
-# Pass files as script parameters or piped as input:
- Do NOT code paths to files in the tree, as this breaks out-of-tree builds.
- If you must, then you must also use an automake rule to create the script.
-# use @c '#!/usr/bin/perl' as the first line of Perl scripts.
-# always <code>use strict</code> and <code>use warnings</code>
-# invoke scripts indirectly in Makefiles or other scripts:
@code
perl script.pl
@endcode
Maintainers must also be sure to follow additional guidelines:
-# Ensure that Perl scripts are committed as executables:
- Use "<code>chmod +x script.pl</code>"
@a before using "<code>svn add script.pl</code>", or
- Use "<code>svn ps svn:executable '*' script.pl</code>"
@a after using "<code>svn add script.pl</code>".
*/
/** @page styleautotools Autotools Style Guide
This page contains style guidelines for the OpenOCD autotools scripts.
The following guidelines apply to the @c configure.in file:
- Better guidelines need to be developed, but until then...
- Use good judgement.
The following guidelines apply to @c Makefile.am files:
-# When assigning variables with long lists of items:
-# Separate the values on each line to make the files "patch friendly":
@code
VAR = \
value1 \
value2 \
...
value9 \
value10
@endcode
*/
/** @file
This file contains the @ref styleguide pages. The @ref styleguide pages
include the following Style Guides for their respective code and
documentation languages:
- @ref styletcl
- @ref stylec
- @ref styledoxygen
- @ref styletexinfo
- @ref stylelatex
- @ref styleperl
- @ref styleautotools
*/