1Like other projects, we also have some guidelines to keep to the 2code. For Git in general, a few 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 - Fixing style violations while working on a real change as a 22 preparatory clean-up step is good, but otherwise avoid useless code 23 churn for the sake of conforming to the style. 24 25 "Once it _is_ in the tree, it's not really worth the patch noise to 26 go and fix it up." 27 Cf. http://article.gmane.org/gmane.linux.kernel/943020 28 29Make your code readable and sensible, and don't try to be clever. 30 31As for more concrete guidelines, just imitate the existing code 32(this is a good guideline, no matter which project you are 33contributing to). It is always preferable to match the _local_ 34convention. New code added to Git suite is expected to match 35the overall style of existing code. Modifications to existing 36code is expected to match the style the surrounding code already 37uses (even if it doesn't match the overall style of existing code). 38 39But if you must have a list of rules, here they are. 40 41For shell scripts specifically (not exhaustive): 42 43 - We use tabs for indentation. 44 45 - Case arms are indented at the same depth as case and esac lines, 46 like this: 47 48 case "$variable" in 49 pattern1) 50 do this 51 ;; 52 pattern2) 53 do that 54 ;; 55 esac 56 57 - Redirection operators should be written with space before, but no 58 space after them. In other words, write 'echo test >"$file"' 59 instead of 'echo test> $file' or 'echo test > $file'. Note that 60 even though it is not required by POSIX to double-quote the 61 redirection target in a variable (as shown above), our code does so 62 because some versions of bash issue a warning without the quotes. 63 64 (incorrect) 65 cat hello > world < universe 66 echo hello >$world 67 68 (correct) 69 cat hello >world <universe 70 echo hello >"$world" 71 72 - We prefer $( ... ) for command substitution; unlike ``, it 73 properly nests. It should have been the way Bourne spelled 74 it from day one, but unfortunately isn't. 75 76 - If you want to find out if a command is available on the user's 77 $PATH, you should use 'type <command>', instead of 'which <command>'. 78 The output of 'which' is not machine parseable and its exit code 79 is not reliable across platforms. 80 81 - We use POSIX compliant parameter substitutions and avoid bashisms; 82 namely: 83 84 - We use ${parameter-word} and its [-=?+] siblings, and their 85 colon'ed "unset or null" form. 86 87 - We use ${parameter#word} and its [#%] siblings, and their 88 doubled "longest matching" form. 89 90 - No "Substring Expansion" ${parameter:offset:length}. 91 92 - No shell arrays. 93 94 - No strlen ${#parameter}. 95 96 - No pattern replacement ${parameter/pattern/string}. 97 98 - We use Arithmetic Expansion $(( ... )). 99 100 - Inside Arithmetic Expansion, spell shell variables with $ in front 101 of them, as some shells do not grok $((x)) while accepting $(($x)) 102 just fine (e.g. dash older than 0.5.4). 103 104 - We do not use Process Substitution <(list) or >(list). 105 106 - Do not write control structures on a single line with semicolon. 107 "then" should be on the next line for if statements, and "do" 108 should be on the next line for "while" and "for". 109 110 (incorrect) 111 if test -f hello; then 112 do this 113 fi 114 115 (correct) 116 if test -f hello 117 then 118 do this 119 fi 120 121 - We prefer "test" over "[ ... ]". 122 123 - We do not write the noiseword "function" in front of shell 124 functions. 125 126 - We prefer a space between the function name and the parentheses, 127 and no space inside the parentheses. The opening "{" should also 128 be on the same line. 129 130 (incorrect) 131 my_function(){ 132 ... 133 134 (correct) 135 my_function () { 136 ... 137 138 - As to use of grep, stick to a subset of BRE (namely, no \{m,n\}, 139 [::], [==], or [..]) for portability. 140 141 - We do not use \{m,n\}; 142 143 - We do not use -E; 144 145 - We do not use ? or + (which are \{0,1\} and \{1,\} 146 respectively in BRE) but that goes without saying as these 147 are ERE elements not BRE (note that \? and \+ are not even part 148 of BRE -- making them accessible from BRE is a GNU extension). 149 150 - Use Git's gettext wrappers in git-sh-i18n to make the user 151 interface translatable. See "Marking strings for translation" in 152 po/README. 153 154 - We do not write our "test" command with "-a" and "-o" and use "&&" 155 or "||" to concatenate multiple "test" commands instead, because 156 the use of "-a/-o" is often error-prone. E.g. 157 158 test -n "$x" -a "$a" = "$b" 159 160 is buggy and breaks when $x is "=", but 161 162 test -n "$x" && test "$a" = "$b" 163 164 does not have such a problem. 165 166 167For C programs: 168 169 - We use tabs to indent, and interpret tabs as taking up to 170 8 spaces. 171 172 - We try to keep to at most 80 characters per line. 173 174 - As a Git developer we assume you have a reasonably modern compiler 175 and we recommend you to enable the DEVELOPER makefile knob to 176 ensure your patch is clear of all compiler warnings we care about, 177 by e.g. "echo DEVELOPER=1 >>config.mak". 178 179 - We try to support a wide range of C compilers to compile Git with, 180 including old ones. That means that you should not use C99 181 initializers, even if a lot of compilers grok it. 182 183 - Variables have to be declared at the beginning of the block. 184 185 - NULL pointers shall be written as NULL, not as 0. 186 187 - When declaring pointers, the star sides with the variable 188 name, i.e. "char *string", not "char* string" or 189 "char * string". This makes it easier to understand code 190 like "char *string, c;". 191 192 - Use whitespace around operators and keywords, but not inside 193 parentheses and not around functions. So: 194 195 while (condition) 196 func(bar + 1); 197 198 and not: 199 200 while( condition ) 201 func (bar+1); 202 203 - We avoid using braces unnecessarily. I.e. 204 205 if (bla) { 206 x = 1; 207 } 208 209 is frowned upon. But there are a few exceptions: 210 211 - When the statement extends over a few lines (e.g., a while loop 212 with an embedded conditional, or a comment). E.g.: 213 214 while (foo) { 215 if (x) 216 one(); 217 else 218 two(); 219 } 220 221 if (foo) { 222 /* 223 * This one requires some explanation, 224 * so we're better off with braces to make 225 * it obvious that the indentation is correct. 226 */ 227 doit(); 228 } 229 230 - When there are multiple arms to a conditional and some of them 231 require braces, enclose even a single line block in braces for 232 consistency. E.g.: 233 234 if (foo) { 235 doit(); 236 } else { 237 one(); 238 two(); 239 three(); 240 } 241 242 - We try to avoid assignments in the condition of an "if" statement. 243 244 - Try to make your code understandable. You may put comments 245 in, but comments invariably tend to stale out when the code 246 they were describing changes. Often splitting a function 247 into two makes the intention of the code much clearer. 248 249 - Multi-line comments include their delimiters on separate lines from 250 the text. E.g. 251 252 /* 253 * A very long 254 * multi-line comment. 255 */ 256 257 Note however that a comment that explains a translatable string to 258 translators uses a convention of starting with a magic token 259 "TRANSLATORS: ", e.g. 260 261 /* 262 * TRANSLATORS: here is a comment that explains the string to 263 * be translated, that follows immediately after it. 264 */ 265 _("Here is a translatable string explained by the above."); 266 267 - Double negation is often harder to understand than no negation 268 at all. 269 270 - There are two schools of thought when it comes to comparison, 271 especially inside a loop. Some people prefer to have the less stable 272 value on the left hand side and the more stable value on the right hand 273 side, e.g. if you have a loop that counts variable i down to the 274 lower bound, 275 276 while (i > lower_bound) { 277 do something; 278 i--; 279 } 280 281 Other people prefer to have the textual order of values match the 282 actual order of values in their comparison, so that they can 283 mentally draw a number line from left to right and place these 284 values in order, i.e. 285 286 while (lower_bound < i) { 287 do something; 288 i--; 289 } 290 291 Both are valid, and we use both. However, the more "stable" the 292 stable side becomes, the more we tend to prefer the former 293 (comparison with a constant, "i > 0", is an extreme example). 294 Just do not mix styles in the same part of the code and mimic 295 existing styles in the neighbourhood. 296 297 - There are two schools of thought when it comes to splitting a long 298 logical line into multiple lines. Some people push the second and 299 subsequent lines far enough to the right with tabs and align them: 300 301 if (the_beginning_of_a_very_long_expression_that_has_to || 302 span_more_than_a_single_line_of || 303 the_source_text) { 304 ... 305 306 while other people prefer to align the second and the subsequent 307 lines with the column immediately inside the opening parenthesis, 308 with tabs and spaces, following our "tabstop is always a multiple 309 of 8" convention: 310 311 if (the_beginning_of_a_very_long_expression_that_has_to || 312 span_more_than_a_single_line_of || 313 the_source_text) { 314 ... 315 316 Both are valid, and we use both. Again, just do not mix styles in 317 the same part of the code and mimic existing styles in the 318 neighbourhood. 319 320 - When splitting a long logical line, some people change line before 321 a binary operator, so that the result looks like a parse tree when 322 you turn your head 90-degrees counterclockwise: 323 324 if (the_beginning_of_a_very_long_expression_that_has_to 325 || span_more_than_a_single_line_of_the_source_text) { 326 327 while other people prefer to leave the operator at the end of the 328 line: 329 330 if (the_beginning_of_a_very_long_expression_that_has_to || 331 span_more_than_a_single_line_of_the_source_text) { 332 333 Both are valid, but we tend to use the latter more, unless the 334 expression gets fairly complex, in which case the former tends to 335 be easier to read. Again, just do not mix styles in the same part 336 of the code and mimic existing styles in the neighbourhood. 337 338 - When splitting a long logical line, with everything else being 339 equal, it is preferable to split after the operator at higher 340 level in the parse tree. That is, this is more preferable: 341 342 if (a_very_long_variable * that_is_used_in + 343 a_very_long_expression) { 344 ... 345 346 than 347 348 if (a_very_long_variable * 349 that_is_used_in + a_very_long_expression) { 350 ... 351 352 - Some clever tricks, like using the !! operator with arithmetic 353 constructs, can be extremely confusing to others. Avoid them, 354 unless there is a compelling reason to use them. 355 356 - Use the API. No, really. We have a strbuf (variable length 357 string), several arrays with the ALLOC_GROW() macro, a 358 string_list for sorted string lists, a hash map (mapping struct 359 objects) named "struct decorate", amongst other things. 360 361 - When you come up with an API, document it. 362 363 - The first #include in C files, except in platform specific compat/ 364 implementations, must be either "git-compat-util.h", "cache.h" or 365 "builtin.h". You do not have to include more than one of these. 366 367 - A C file must directly include the header files that declare the 368 functions and the types it uses, except for the functions and types 369 that are made available to it by including one of the header files 370 it must include by the previous rule. 371 372 - If you are planning a new command, consider writing it in shell 373 or perl first, so that changes in semantics can be easily 374 changed and discussed. Many Git commands started out like 375 that, and a few are still scripts. 376 377 - Avoid introducing a new dependency into Git. This means you 378 usually should stay away from scripting languages not already 379 used in the Git core command set (unless your command is clearly 380 separate from it, such as an importer to convert random-scm-X 381 repositories to Git). 382 383 - When we pass <string, length> pair to functions, we should try to 384 pass them in that order. 385 386 - Use Git's gettext wrappers to make the user interface 387 translatable. See "Marking strings for translation" in po/README. 388 389For Perl programs: 390 391 - Most of the C guidelines above apply. 392 393 - We try to support Perl 5.8 and later ("use Perl 5.008"). 394 395 - use strict and use warnings are strongly preferred. 396 397 - Don't overuse statement modifiers unless using them makes the 398 result easier to follow. 399 400 ... do something ... 401 do_this() unless (condition); 402 ... do something else ... 403 404 is more readable than: 405 406 ... do something ... 407 unless (condition) { 408 do_this(); 409 } 410 ... do something else ... 411 412 *only* when the condition is so rare that do_this() will be almost 413 always called. 414 415 - We try to avoid assignments inside "if ()" conditions. 416 417 - Learn and use Git.pm if you need that functionality. 418 419 - For Emacs, it's useful to put the following in 420 GIT_CHECKOUT/.dir-locals.el, assuming you use cperl-mode: 421 422 ;; note the first part is useful for C editing, too 423 ((nil . ((indent-tabs-mode . t) 424 (tab-width . 8) 425 (fill-column . 80))) 426 (cperl-mode . ((cperl-indent-level . 8) 427 (cperl-extra-newline-before-brace . nil) 428 (cperl-merge-trailing-else . t)))) 429 430For Python scripts: 431 432 - We follow PEP-8 (http://www.python.org/dev/peps/pep-0008/). 433 434 - As a minimum, we aim to be compatible with Python 2.6 and 2.7. 435 436 - Where required libraries do not restrict us to Python 2, we try to 437 also be compatible with Python 3.1 and later. 438 439 - When you must differentiate between Unicode literals and byte string 440 literals, it is OK to use the 'b' prefix. Even though the Python 441 documentation for version 2.6 does not mention this prefix, it has 442 been supported since version 2.6.0. 443 444Error Messages 445 446 - Do not end error messages with a full stop. 447 448 - Do not capitalize ("unable to open %s", not "Unable to open %s") 449 450 - Say what the error is first ("cannot open %s", not "%s: cannot open") 451 452 453Externally Visible Names 454 455 - For configuration variable names, follow the existing convention: 456 457 . The section name indicates the affected subsystem. 458 459 . The subsection name, if any, indicates which of an unbounded set 460 of things to set the value for. 461 462 . The variable name describes the effect of tweaking this knob. 463 464 The section and variable names that consist of multiple words are 465 formed by concatenating the words without punctuations (e.g. `-`), 466 and are broken using bumpyCaps in documentation as a hint to the 467 reader. 468 469 When choosing the variable namespace, do not use variable name for 470 specifying possibly unbounded set of things, most notably anything 471 an end user can freely come up with (e.g. branch names). Instead, 472 use subsection names or variable values, like the existing variable 473 branch.<name>.description does. 474 475 476Writing Documentation: 477 478 Most (if not all) of the documentation pages are written in the 479 AsciiDoc format in *.txt files (e.g. Documentation/git.txt), and 480 processed into HTML and manpages (e.g. git.html and git.1 in the 481 same directory). 482 483 The documentation liberally mixes US and UK English (en_US/UK) 484 norms for spelling and grammar, which is somewhat unfortunate. 485 In an ideal world, it would have been better if it consistently 486 used only one and not the other, and we would have picked en_US 487 (if you wish to correct the English of some of the existing 488 documentation, please see the documentation-related advice in the 489 Documentation/SubmittingPatches file). 490 491 Every user-visible change should be reflected in the documentation. 492 The same general rule as for code applies -- imitate the existing 493 conventions. 494 495 A few commented examples follow to provide reference when writing or 496 modifying command usage strings and synopsis sections in the manual 497 pages: 498 499 Placeholders are spelled in lowercase and enclosed in angle brackets: 500 <file> 501 --sort=<key> 502 --abbrev[=<n>] 503 504 If a placeholder has multiple words, they are separated by dashes: 505 <new-branch-name> 506 --template=<template-directory> 507 508 Possibility of multiple occurrences is indicated by three dots: 509 <file>... 510 (One or more of <file>.) 511 512 Optional parts are enclosed in square brackets: 513 [<extra>] 514 (Zero or one <extra>.) 515 516 --exec-path[=<path>] 517 (Option with an optional argument. Note that the "=" is inside the 518 brackets.) 519 520 [<patch>...] 521 (Zero or more of <patch>. Note that the dots are inside, not 522 outside the brackets.) 523 524 Multiple alternatives are indicated with vertical bars: 525 [-q | --quiet] 526 [--utf8 | --no-utf8] 527 528 Parentheses are used for grouping: 529 [(<rev> | <range>)...] 530 (Any number of either <rev> or <range>. Parens are needed to make 531 it clear that "..." pertains to both <rev> and <range>.) 532 533 [(-p <parent>)...] 534 (Any number of option -p, each with one <parent> argument.) 535 536 git remote set-head <name> (-a | -d | <branch>) 537 (One and only one of "-a", "-d" or "<branch>" _must_ (no square 538 brackets) be provided.) 539 540 And a somewhat more contrived example: 541 --diff-filter=[(A|C|D|M|R|T|U|X|B)...[*]] 542 Here "=" is outside the brackets, because "--diff-filter=" is a 543 valid usage. "*" has its own pair of brackets, because it can 544 (optionally) be specified only when one or more of the letters is 545 also provided. 546 547 A note on notation: 548 Use 'git' (all lowercase) when talking about commands i.e. something 549 the user would type into a shell and use 'Git' (uppercase first letter) 550 when talking about the version control system and its properties. 551 552 A few commented examples follow to provide reference when writing or 553 modifying paragraphs or option/command explanations that contain options 554 or commands: 555 556 Literal examples (e.g. use of command-line options, command names, 557 branch names, configuration and environment variables) must be 558 typeset in monospace (i.e. wrapped with backticks): 559 `--pretty=oneline` 560 `git rev-list` 561 `remote.pushDefault` 562 `GIT_DIR` 563 `HEAD` 564 565 An environment variable must be prefixed with "$" only when referring to its 566 value and not when referring to the variable itself, in this case there is 567 nothing to add except the backticks: 568 `GIT_DIR` is specified 569 `$GIT_DIR/hooks/pre-receive` 570 571 Word phrases enclosed in `backtick characters` are rendered literally 572 and will not be further expanded. The use of `backticks` to achieve the 573 previous rule means that literal examples should not use AsciiDoc 574 escapes. 575 Correct: 576 `--pretty=oneline` 577 Incorrect: 578 `\--pretty=oneline` 579 580 If some place in the documentation needs to typeset a command usage 581 example with inline substitutions, it is fine to use +monospaced and 582 inline substituted text+ instead of `monospaced literal text`, and with 583 the former, the part that should not get substituted must be 584 quoted/escaped.