SubmittingPatches: final submission is To: maintainer and CC: list
[git/git.git] / Documentation / CodingGuidelines
1 Like other projects, we also have some guidelines to keep to the
2 code. For Git in general, three rough rules are:
3
4 - Most importantly, we never say "It's in POSIX; we'll happily
5 ignore your needs should your system not conform to it."
6 We live in the real world.
7
8 - However, we often say "Let's stay away from that construct,
9 it's not even in POSIX".
10
11 - In spite of the above two rules, we sometimes say "Although
12 this is not in POSIX, it (is so convenient | makes the code
13 much more readable | has other good characteristics) and
14 practically all the platforms we care about support it, so
15 let's use it".
16
17 Again, we live in the real world, and it is sometimes a
18 judgement call, the decision based more on real world
19 constraints people face than what the paper standard says.
20
21 Make your code readable and sensible, and don't try to be clever.
22
23 As for more concrete guidelines, just imitate the existing code
24 (this is a good guideline, no matter which project you are
25 contributing to). It is always preferable to match the _local_
26 convention. New code added to Git suite is expected to match
27 the overall style of existing code. Modifications to existing
28 code is expected to match the style the surrounding code already
29 uses (even if it doesn't match the overall style of existing code).
30
31 But if you must have a list of rules, here they are.
32
33 For shell scripts specifically (not exhaustive):
34
35 - We use tabs for indentation.
36
37 - Case arms are indented at the same depth as case and esac lines.
38
39 - Redirection operators should be written with space before, but no
40 space after them. In other words, write 'echo test >"$file"'
41 instead of 'echo test> $file' or 'echo test > $file'. Note that
42 even though it is not required by POSIX to double-quote the
43 redirection target in a variable (as shown above), our code does so
44 because some versions of bash issue a warning without the quotes.
45
46 - We prefer $( ... ) for command substitution; unlike ``, it
47 properly nests. It should have been the way Bourne spelled
48 it from day one, but unfortunately isn't.
49
50 - If you want to find out if a command is available on the user's
51 $PATH, you should use 'type <command>', instead of 'which <command>'.
52 The output of 'which' is not machine parseable and its exit code
53 is not reliable across platforms.
54
55 - We use POSIX compliant parameter substitutions and avoid bashisms;
56 namely:
57
58 - We use ${parameter-word} and its [-=?+] siblings, and their
59 colon'ed "unset or null" form.
60
61 - We use ${parameter#word} and its [#%] siblings, and their
62 doubled "longest matching" form.
63
64 - No "Substring Expansion" ${parameter:offset:length}.
65
66 - No shell arrays.
67
68 - No strlen ${#parameter}.
69
70 - No pattern replacement ${parameter/pattern/string}.
71
72 - We use Arithmetic Expansion $(( ... )).
73
74 - Inside Arithmetic Expansion, spell shell variables with $ in front
75 of them, as some shells do not grok $((x)) while accepting $(($x))
76 just fine (e.g. dash older than 0.5.4).
77
78 - We do not use Process Substitution <(list) or >(list).
79
80 - Do not write control structures on a single line with semicolon.
81 "then" should be on the next line for if statements, and "do"
82 should be on the next line for "while" and "for".
83
84 - We prefer "test" over "[ ... ]".
85
86 - We do not write the noiseword "function" in front of shell
87 functions.
88
89 - We prefer a space between the function name and the parentheses. The
90 opening "{" should also be on the same line.
91 E.g.: my_function () {
92
93 - As to use of grep, stick to a subset of BRE (namely, no \{m,n\},
94 [::], [==], nor [..]) for portability.
95
96 - We do not use \{m,n\};
97
98 - We do not use -E;
99
100 - We do not use ? nor + (which are \{0,1\} and \{1,\}
101 respectively in BRE) but that goes without saying as these
102 are ERE elements not BRE (note that \? and \+ are not even part
103 of BRE -- making them accessible from BRE is a GNU extension).
104
105 - Use Git's gettext wrappers in git-sh-i18n to make the user
106 interface translatable. See "Marking strings for translation" in
107 po/README.
108
109 For C programs:
110
111 - We use tabs to indent, and interpret tabs as taking up to
112 8 spaces.
113
114 - We try to keep to at most 80 characters per line.
115
116 - We try to support a wide range of C compilers to compile Git with,
117 including old ones. That means that you should not use C99
118 initializers, even if a lot of compilers grok it.
119
120 - Variables have to be declared at the beginning of the block.
121
122 - NULL pointers shall be written as NULL, not as 0.
123
124 - When declaring pointers, the star sides with the variable
125 name, i.e. "char *string", not "char* string" or
126 "char * string". This makes it easier to understand code
127 like "char *string, c;".
128
129 - We avoid using braces unnecessarily. I.e.
130
131 if (bla) {
132 x = 1;
133 }
134
135 is frowned upon. A gray area is when the statement extends
136 over a few lines, and/or you have a lengthy comment atop of
137 it. Also, like in the Linux kernel, if there is a long list
138 of "else if" statements, it can make sense to add braces to
139 single line blocks.
140
141 - We try to avoid assignments inside if().
142
143 - Try to make your code understandable. You may put comments
144 in, but comments invariably tend to stale out when the code
145 they were describing changes. Often splitting a function
146 into two makes the intention of the code much clearer.
147
148 - Multi-line comments include their delimiters on separate lines from
149 the text. E.g.
150
151 /*
152 * A very long
153 * multi-line comment.
154 */
155
156 - Double negation is often harder to understand than no negation
157 at all.
158
159 - Some clever tricks, like using the !! operator with arithmetic
160 constructs, can be extremely confusing to others. Avoid them,
161 unless there is a compelling reason to use them.
162
163 - Use the API. No, really. We have a strbuf (variable length
164 string), several arrays with the ALLOC_GROW() macro, a
165 string_list for sorted string lists, a hash map (mapping struct
166 objects) named "struct decorate", amongst other things.
167
168 - When you come up with an API, document it.
169
170 - The first #include in C files, except in platform specific
171 compat/ implementations, should be git-compat-util.h or another
172 header file that includes it, such as cache.h or builtin.h.
173
174 - If you are planning a new command, consider writing it in shell
175 or perl first, so that changes in semantics can be easily
176 changed and discussed. Many Git commands started out like
177 that, and a few are still scripts.
178
179 - Avoid introducing a new dependency into Git. This means you
180 usually should stay away from scripting languages not already
181 used in the Git core command set (unless your command is clearly
182 separate from it, such as an importer to convert random-scm-X
183 repositories to Git).
184
185 - When we pass <string, length> pair to functions, we should try to
186 pass them in that order.
187
188 - Use Git's gettext wrappers to make the user interface
189 translatable. See "Marking strings for translation" in po/README.
190
191 For Perl programs:
192
193 - Most of the C guidelines above apply.
194
195 - We try to support Perl 5.8 and later ("use Perl 5.008").
196
197 - use strict and use warnings are strongly preferred.
198
199 - Don't overuse statement modifiers unless using them makes the
200 result easier to follow.
201
202 ... do something ...
203 do_this() unless (condition);
204 ... do something else ...
205
206 is more readable than:
207
208 ... do something ...
209 unless (condition) {
210 do_this();
211 }
212 ... do something else ...
213
214 *only* when the condition is so rare that do_this() will be almost
215 always called.
216
217 - We try to avoid assignments inside "if ()" conditions.
218
219 - Learn and use Git.pm if you need that functionality.
220
221 - For Emacs, it's useful to put the following in
222 GIT_CHECKOUT/.dir-locals.el, assuming you use cperl-mode:
223
224 ;; note the first part is useful for C editing, too
225 ((nil . ((indent-tabs-mode . t)
226 (tab-width . 8)
227 (fill-column . 80)))
228 (cperl-mode . ((cperl-indent-level . 8)
229 (cperl-extra-newline-before-brace . nil)
230 (cperl-merge-trailing-else . t))))
231
232 For Python scripts:
233
234 - We follow PEP-8 (http://www.python.org/dev/peps/pep-0008/).
235
236 - As a minimum, we aim to be compatible with Python 2.6 and 2.7.
237
238 - Where required libraries do not restrict us to Python 2, we try to
239 also be compatible with Python 3.1 and later.
240
241 - When you must differentiate between Unicode literals and byte string
242 literals, it is OK to use the 'b' prefix. Even though the Python
243 documentation for version 2.6 does not mention this prefix, it has
244 been supported since version 2.6.0.
245
246 Writing Documentation:
247
248 Most (if not all) of the documentation pages are written in the
249 AsciiDoc format in *.txt files (e.g. Documentation/git.txt), and
250 processed into HTML and manpages (e.g. git.html and git.1 in the
251 same directory).
252
253 The documentation liberally mixes US and UK English (en_US/UK)
254 norms for spelling and grammar, which is somewhat unfortunate.
255 In an ideal world, it would have been better if it consistently
256 used only one and not the other, and we would have picked en_US
257 (if you wish to correct the English of some of the existing
258 documentation, please see the documentation-related advice in the
259 Documentation/SubmittingPatches file).
260
261 Every user-visible change should be reflected in the documentation.
262 The same general rule as for code applies -- imitate the existing
263 conventions.
264
265 A few commented examples follow to provide reference when writing or
266 modifying command usage strings and synopsis sections in the manual
267 pages:
268
269 Placeholders are spelled in lowercase and enclosed in angle brackets:
270 <file>
271 --sort=<key>
272 --abbrev[=<n>]
273
274 Possibility of multiple occurrences is indicated by three dots:
275 <file>...
276 (One or more of <file>.)
277
278 Optional parts are enclosed in square brackets:
279 [<extra>]
280 (Zero or one <extra>.)
281
282 --exec-path[=<path>]
283 (Option with an optional argument. Note that the "=" is inside the
284 brackets.)
285
286 [<patch>...]
287 (Zero or more of <patch>. Note that the dots are inside, not
288 outside the brackets.)
289
290 Multiple alternatives are indicated with vertical bar:
291 [-q | --quiet]
292 [--utf8 | --no-utf8]
293
294 Parentheses are used for grouping:
295 [(<rev>|<range>)...]
296 (Any number of either <rev> or <range>. Parens are needed to make
297 it clear that "..." pertains to both <rev> and <range>.)
298
299 [(-p <parent>)...]
300 (Any number of option -p, each with one <parent> argument.)
301
302 git remote set-head <name> (-a | -d | <branch>)
303 (One and only one of "-a", "-d" or "<branch>" _must_ (no square
304 brackets) be provided.)
305
306 And a somewhat more contrived example:
307 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]]
308 Here "=" is outside the brackets, because "--diff-filter=" is a
309 valid usage. "*" has its own pair of brackets, because it can
310 (optionally) be specified only when one or more of the letters is
311 also provided.
312
313 A note on notation:
314 Use 'git' (all lowercase) when talking about commands i.e. something
315 the user would type into a shell and use 'Git' (uppercase first letter)
316 when talking about the version control system and its properties.
317
318 A few commented examples follow to provide reference when writing or
319 modifying paragraphs or option/command explanations that contain options
320 or commands:
321
322 Literal examples (e.g. use of command-line options, command names, and
323 configuration variables) are typeset in monospace, and if you can use
324 `backticks around word phrases`, do so.
325 `--pretty=oneline`
326 `git rev-list`
327 `remote.pushdefault`
328
329 Word phrases enclosed in `backtick characters` are rendered literally
330 and will not be further expanded. The use of `backticks` to achieve the
331 previous rule means that literal examples should not use AsciiDoc
332 escapes.
333 Correct:
334 `--pretty=oneline`
335 Incorrect:
336 `\--pretty=oneline`
337
338 If some place in the documentation needs to typeset a command usage
339 example with inline substitutions, it is fine to use +monospaced and
340 inline substituted text+ instead of `monospaced literal text`, and with
341 the former, the part that should not get substituted must be
342 quoted/escaped.