| 1 | ---
|
| 2 | title: Builtin Commands (Oils Reference)
|
| 3 | all_docs_url: ..
|
| 4 | body_css_class: width40
|
| 5 | default_highlighter: oils-sh
|
| 6 | preserve_anchor_case: yes
|
| 7 | ---
|
| 8 |
|
| 9 | <div class="doc-ref-header">
|
| 10 |
|
| 11 | [Oils Reference](index.html) — Chapter **Builtin Commands**
|
| 12 |
|
| 13 | </div>
|
| 14 |
|
| 15 | This chapter in the [Oils Reference](index.html) describes builtin commands for OSH and YSH.
|
| 16 |
|
| 17 | <span class="in-progress">(in progress)</span>
|
| 18 |
|
| 19 | <div id="dense-toc">
|
| 20 | </div>
|
| 21 |
|
| 22 | ## Memory
|
| 23 |
|
| 24 | ### cmd/append
|
| 25 |
|
| 26 | Append word arguments to a list:
|
| 27 |
|
| 28 | var mylist = :| hello |
|
| 29 |
|
| 30 | append *.py (mylist) # append all Python files
|
| 31 |
|
| 32 | var myflags = []
|
| 33 | append -- -c 'echo hi' (myflags) # -- to avoid ambiguity
|
| 34 |
|
| 35 | It's a shortcut for:
|
| 36 |
|
| 37 | call myflags->append('-c')
|
| 38 | call myflags->append('echo hi')
|
| 39 |
|
| 40 | Similar names: [append][]
|
| 41 |
|
| 42 | [append]: chap-index.html#append
|
| 43 |
|
| 44 | ### pp
|
| 45 |
|
| 46 | The `pp` builtin pretty prints values and interpreter state.
|
| 47 |
|
| 48 | Pretty printing expressions is the most common:
|
| 49 |
|
| 50 | $ var x = 42
|
| 51 | $ pp (x + 5)
|
| 52 | myfile.ysh:1: (Int) 47 # print value with code location
|
| 53 |
|
| 54 | You can pass an unevaluated expression:
|
| 55 |
|
| 56 | $ pp [x + 5]
|
| 57 | myfile.ysh:1: (Int) 47 # evaluate first
|
| 58 |
|
| 59 | The `value` command is a synonym for the interactive `=` operator:
|
| 60 |
|
| 61 | $ pp value (x)
|
| 62 | (Int) 42
|
| 63 |
|
| 64 | $ = x
|
| 65 | (Int) 42
|
| 66 |
|
| 67 | Print proc names and doc comments:
|
| 68 |
|
| 69 | $ pp proc # subject to change
|
| 70 |
|
| 71 | You can also print low-level interpreter state. The trailing `_` indicates
|
| 72 | that the exact format may change:
|
| 73 |
|
| 74 | Examples:
|
| 75 |
|
| 76 | $ var x = :| one two |
|
| 77 |
|
| 78 | $ pp asdl_ (x) # dump the ASDL "guts"
|
| 79 |
|
| 80 | $ pp test_ (x) # single-line stable format, for spec tests
|
| 81 |
|
| 82 | # dump the ASDL representation of a "Cell", which is a location for a value
|
| 83 | # (not the value itself)
|
| 84 | $ pp cell_ x
|
| 85 |
|
| 86 |
|
| 87 | ## Handle Errors
|
| 88 |
|
| 89 | ### error
|
| 90 |
|
| 91 | The `error` builtin interrupts shell execution.
|
| 92 |
|
| 93 | If there's a surrounding `try` block, the `_error` register is set, and
|
| 94 | execution proceeds after the block.
|
| 95 |
|
| 96 | Otherwise, the shell exits with a non-zero status.
|
| 97 |
|
| 98 | Examples:
|
| 99 |
|
| 100 | error 'Missing /tmp' # program fails with status 10
|
| 101 |
|
| 102 | try {
|
| 103 | error 'Another problem'
|
| 104 | }
|
| 105 | echo $[error.code] # => 10
|
| 106 |
|
| 107 | Override the default error code of `10` with a named argument:
|
| 108 |
|
| 109 | error 'Missing /tmp' (code=99) # program fails with status 99
|
| 110 |
|
| 111 | Named arguments add arbitrary properties to the resulting `_error` register:
|
| 112 |
|
| 113 | error 'Oops' (path='foo.json')
|
| 114 |
|
| 115 | See [YSH Error Handling](../ysh-error-handling.html) for more examples.
|
| 116 |
|
| 117 | ### failed
|
| 118 |
|
| 119 | A shortcut for `(_error.code !== 0)`:
|
| 120 |
|
| 121 | try {
|
| 122 | ls /tmp
|
| 123 | }
|
| 124 | if failed {
|
| 125 | echo 'ls failed'
|
| 126 | }
|
| 127 |
|
| 128 | It saves you 7 punctuation characters: `( _ . !== )`
|
| 129 |
|
| 130 | See [YSH Error Handling](../ysh-error-handling.html) for more examples.
|
| 131 |
|
| 132 | ### try
|
| 133 |
|
| 134 | Run a block of code, stopping at the first error. (This is implemented with
|
| 135 | `shopt --set errexit`)
|
| 136 |
|
| 137 | `try` sets the `_error` register to a dict, and always returns 0.
|
| 138 |
|
| 139 | try {
|
| 140 | ls /nonexistent
|
| 141 | }
|
| 142 | if (_error.code !== 0) {
|
| 143 | echo 'ls failed'
|
| 144 | }
|
| 145 |
|
| 146 | Handle expression errors:
|
| 147 |
|
| 148 | try {
|
| 149 | var x = 42 / 0
|
| 150 | }
|
| 151 |
|
| 152 | And errors from compound commands:
|
| 153 |
|
| 154 | try {
|
| 155 | ls | wc -l
|
| 156 | diff <(sort left.txt) <(sort right.txt)
|
| 157 | }
|
| 158 |
|
| 159 | The case statement can be useful:
|
| 160 |
|
| 161 | try {
|
| 162 | grep PATTERN FILE.txt
|
| 163 | }
|
| 164 | case (_error.code) {
|
| 165 | (0) { echo 'found' }
|
| 166 | (1) { echo 'not found' }
|
| 167 | (else) { echo "grep returned status $[_error.code]" }
|
| 168 | }
|
| 169 |
|
| 170 | See [YSH Error Handling](../ysh-error-handling.html) for more examples.
|
| 171 |
|
| 172 | ### boolstatus
|
| 173 |
|
| 174 | Runs a command, and requires the exit code to be 0 or 1.
|
| 175 |
|
| 176 | if boolstatus egrep '[0-9]+' myfile { # e.g. aborts on status 2
|
| 177 | echo 'found' # status 0 means found
|
| 178 | } else {
|
| 179 | echo 'not found' # status 1 means not found
|
| 180 | }
|
| 181 |
|
| 182 | It's meant for external commands that "return" more than 2 values, like true /
|
| 183 | false / fail, rather than pass / fail.
|
| 184 |
|
| 185 | ### assert
|
| 186 |
|
| 187 | Evaluates and expression, and fails if it is not truthy.
|
| 188 |
|
| 189 | assert (false) # fails
|
| 190 | assert [false] # also fails (the expression is evaluated)
|
| 191 |
|
| 192 | It's common to pass an unevaluated expression with `===`:
|
| 193 |
|
| 194 | func f() { return (42) }
|
| 195 |
|
| 196 | assert [43 === f()]
|
| 197 |
|
| 198 | In this special case, you get a nicer error message:
|
| 199 |
|
| 200 | > Expected: 43
|
| 201 | > Got: 42
|
| 202 |
|
| 203 | That is, the left-hand side should be the expected value, and the right-hand
|
| 204 | side should be the actual value.
|
| 205 |
|
| 206 | ## Shell State
|
| 207 |
|
| 208 | ### ysh-cd
|
| 209 |
|
| 210 | It takes a block:
|
| 211 |
|
| 212 | cd / {
|
| 213 | echo $PWD
|
| 214 | }
|
| 215 |
|
| 216 | ### ysh-shopt
|
| 217 |
|
| 218 | Sets shell options, e.g.
|
| 219 |
|
| 220 | shopt --unset errexit
|
| 221 | shopt --set errexit
|
| 222 |
|
| 223 | You can set or unset multiple options with the groups `strict:all`,
|
| 224 | `ysh:upgrade`, and `ysh:all`. Example:
|
| 225 |
|
| 226 | shopt --set ysh:upgrade
|
| 227 |
|
| 228 | If a block is passed, then:
|
| 229 |
|
| 230 | 1. the mutated options are pushed onto a stack
|
| 231 | 2. the block is executed
|
| 232 | 3. the options are restored to their original state (even if the block fails to
|
| 233 | execute)
|
| 234 |
|
| 235 | Example:
|
| 236 |
|
| 237 | shopt --unset errexit {
|
| 238 | false
|
| 239 | echo 'ok'
|
| 240 | }
|
| 241 |
|
| 242 | Note that setting `ysh:upgrade` or `ysh:all` may initialize the [ENV][] dict.
|
| 243 |
|
| 244 | Related: [shopt](#shopt)
|
| 245 |
|
| 246 | [ENV]: chap-special-var.html#ENV
|
| 247 |
|
| 248 | ### shvar
|
| 249 |
|
| 250 | Execute a block with a global variable set.
|
| 251 |
|
| 252 | shvar IFS=/ {
|
| 253 | echo "ifs is $IFS"
|
| 254 | }
|
| 255 | echo "ifs restored to $IFS"
|
| 256 |
|
| 257 | ### ctx
|
| 258 |
|
| 259 | Execute a block with a shared "context" that can be updated using the `ctx`
|
| 260 | built-in.
|
| 261 |
|
| 262 | var mydict = {}
|
| 263 | ctx push (mydict) {
|
| 264 | # = mydict => {}
|
| 265 | ctx set (mykey='myval')
|
| 266 | }
|
| 267 | # = mydict => { mykey: 'myval' }
|
| 268 |
|
| 269 | The context can be modified with `ctx set (key=val)`, which updates or inserts
|
| 270 | the value at the given key.
|
| 271 |
|
| 272 | The context can also be updated with `ctx emit field (value)`.
|
| 273 |
|
| 274 | ctx push (mydict) {
|
| 275 | # = mydict => {}
|
| 276 | ctx emit mylist (0)
|
| 277 | # = mydict => { mylist: [0] }
|
| 278 | ctx emit mylist (1)
|
| 279 | }
|
| 280 | # = mydict => { mylist: [0, 1] }
|
| 281 |
|
| 282 | Contexts can be nested, resulting in a stack of contexts.
|
| 283 |
|
| 284 | ctx push (mydict1) {
|
| 285 | ctx set (dict=1)
|
| 286 | ctx push (mydict2) {
|
| 287 | ctx set (dict=2)
|
| 288 | }
|
| 289 | }
|
| 290 | # = mydict1 => { dict: 1 }
|
| 291 | # = mydict2 => { dict: 2 }
|
| 292 |
|
| 293 | `ctx` is useful for creating DSLs, such as a mini-parseArgs.
|
| 294 |
|
| 295 | proc parser (; place ; ; block_def) {
|
| 296 | var p = {}
|
| 297 | ctx push (p, block_def)
|
| 298 | call place->setValue(p)
|
| 299 | }
|
| 300 |
|
| 301 | proc flag (short_name, long_name; type; help) {
|
| 302 | ctx emit flag ({short_name, long_name, type, help})
|
| 303 | }
|
| 304 |
|
| 305 | proc arg (name) {
|
| 306 | ctx emit arg ({name})
|
| 307 | }
|
| 308 |
|
| 309 | parser (&spec) {
|
| 310 | flag -t --tsv (Bool, help='Output as TSV')
|
| 311 | flag -r --recursive (Bool, help='Recurse into the given directory')
|
| 312 | flag -N --count (Int, help='Process no more than N files')
|
| 313 | arg path
|
| 314 | }
|
| 315 |
|
| 316 | ### push-registers
|
| 317 |
|
| 318 | Save global registers like $? on a stack. It's useful for preventing plugins
|
| 319 | from interfering with user code. Example:
|
| 320 |
|
| 321 | status_42 # returns 42 and sets $?
|
| 322 | push-registers { # push a new frame
|
| 323 | status_43 # top of stack changed here
|
| 324 | echo done
|
| 325 | } # stack popped
|
| 326 | echo $? # 42, read from new top-of-stack
|
| 327 |
|
| 328 | Current list of registers:
|
| 329 |
|
| 330 | Regex data underlying BASH_REMATCH, _group(), _start(), _end()
|
| 331 | $?
|
| 332 | _error # set by the try builtin
|
| 333 | PIPESTATUS # aka _pipeline_status
|
| 334 | _process_sub_status
|
| 335 |
|
| 336 |
|
| 337 | ## Modules
|
| 338 |
|
| 339 | ### source-guard
|
| 340 |
|
| 341 | Registers a name in the global "module" dict. Returns 0 if it doesn't exist,
|
| 342 | or 1 if it does.
|
| 343 |
|
| 344 | Use it like this in executable files:
|
| 345 |
|
| 346 | source-guard main || return 0
|
| 347 |
|
| 348 | And like this in libraries:
|
| 349 |
|
| 350 | source-guard myfile.ysh || return 0
|
| 351 |
|
| 352 | ### is-main
|
| 353 |
|
| 354 | The `is-main` builtin returns 1 (false) if the current file was executed with
|
| 355 | the `source` builtin.
|
| 356 |
|
| 357 | In the "main" file, including `-c` or `stdin` input, it returns 0 (true).
|
| 358 |
|
| 359 | Use it like this:
|
| 360 |
|
| 361 | if is-main {
|
| 362 | runproc @ARGV
|
| 363 | }
|
| 364 |
|
| 365 | ### use
|
| 366 |
|
| 367 | The `use` builtin evaluates a source file in a new `Frame`, and then creates an
|
| 368 | `Obj` that is a namespace.
|
| 369 |
|
| 370 | use my-dir/mymodule.ysh
|
| 371 |
|
| 372 | echo $[mymodule.my_integer] # the module Obj has attributes
|
| 373 | mymodule my-proc # the module Obj is invokable
|
| 374 |
|
| 375 | The evaluation of such files is cached, so it won't be re-evaluated if `use` is
|
| 376 | called again.
|
| 377 |
|
| 378 | To import a specific name, use the `--pick` flag:
|
| 379 |
|
| 380 | use my-dir/mymodule.ysh --pick my-proc other-proc
|
| 381 |
|
| 382 | my-proc 1 2
|
| 383 | other-proc 3 4
|
| 384 |
|
| 385 | Note: the `--pick` flag must come *after* the module, so this isn't valid:
|
| 386 |
|
| 387 | use --pick my-proc mymodule.sh # INVALID
|
| 388 |
|
| 389 | <!--
|
| 390 | # TODO:
|
| 391 |
|
| 392 | use mod.ysh --all-provided # relies on __provide__ or provide builtin
|
| 393 | use mod.ysh --all-for-testing
|
| 394 | -->
|
| 395 |
|
| 396 | ---
|
| 397 |
|
| 398 | The `--extern` flag means that `use` does nothing. These commands can be used
|
| 399 | by tools to analyze names.
|
| 400 |
|
| 401 | use --extern grep sed awk
|
| 402 |
|
| 403 | ---
|
| 404 |
|
| 405 | Notes:
|
| 406 |
|
| 407 | - To get a reference to `module-with-hyphens`, you may need to use
|
| 408 | `getVar('module-with-hyphens')`.
|
| 409 | - TODO: consider backtick syntax as well
|
| 410 | - `use` must be used at the top level, not within a function.
|
| 411 | - This behavior is unlike Python.
|
| 412 | - The `use` builtin populates the new module with references to these values in
|
| 413 | the calling module:
|
| 414 | - [ENV][] - to mutate and set environment vars
|
| 415 | - [PS4][] - for cross-module tracing in OSH
|
| 416 |
|
| 417 | [ENV]: chap-special-var.html#ENV
|
| 418 | [PS4]: chap-plugin.html#PS4
|
| 419 |
|
| 420 | Warnings:
|
| 421 |
|
| 422 | - `use` **copies** the module bindings into a new `Obj`. This means that if
|
| 423 | you rebind `mymodule.my_integer`, it will **not** be visible to code in the
|
| 424 | module.
|
| 425 | - This behavior is unlike Python.
|
| 426 | - `use` allows "circular imports". That is `A.ysh` can `use B.ysh`, and vice
|
| 427 | versa.
|
| 428 | - To eliminate confusion over uninitialized names, use **only** `const`,
|
| 429 | `func`, and `proc` at the top level of `my-module.ysh`. Don't run
|
| 430 | commands, use `setvar`, etc.
|
| 431 |
|
| 432 | ## I/O
|
| 433 |
|
| 434 | ### ysh-read
|
| 435 |
|
| 436 | YSH adds long flags to shell's `read`. These two flags are fast and
|
| 437 | recommended:
|
| 438 |
|
| 439 | read --all # whole file including trailing \n, fills $_reply
|
| 440 | read --all (&x) # fills $x
|
| 441 |
|
| 442 | read --num-bytes 3 # read N bytes, fills _reply
|
| 443 | read --num-bytes 3 (&x) # fills $x
|
| 444 |
|
| 445 | ---
|
| 446 |
|
| 447 | This flag replaces shell's `IFS= read -r` idiom, reading one byte a time in an
|
| 448 | unbuffered fashion:
|
| 449 |
|
| 450 | read --raw-line # unbuffered read of line, omitting trailing \n
|
| 451 | read --raw-line (&x) # fills $x
|
| 452 |
|
| 453 | read --raw-line --with-eol # include the trailing \n
|
| 454 |
|
| 455 | A loop over [io.stdin][] allows buffered reading of lines, which is faster.
|
| 456 |
|
| 457 | [io.stdin]: chap-type-method.html#stdin
|
| 458 |
|
| 459 | You may want to use `fromJson8()` or `fromJson()` after reading a line.
|
| 460 |
|
| 461 | ---
|
| 462 |
|
| 463 | The `-0` flag also reads one byte at a time:
|
| 464 |
|
| 465 | read -0 # read until NUL, synonym for read -r -d ''
|
| 466 |
|
| 467 | Notes:
|
| 468 |
|
| 469 | - Unlike OSH [read](#read), none of these features remove NUL bytes.
|
| 470 | - Performance summary: [YSH Input/Output > Three Types of I/O][ysh-io-three]
|
| 471 |
|
| 472 | [ysh-io-three]: ../ysh-io.html#three-types-of-io
|
| 473 |
|
| 474 | <!--
|
| 475 |
|
| 476 | TODO:
|
| 477 |
|
| 478 | - read --netstr
|
| 479 | - io.stdin0 coudl be a buffered version of read -0 ?
|
| 480 | - JSON
|
| 481 | - @() is related - it reads J8 lines
|
| 482 | - JSON lines support?
|
| 483 | - fromJ8Line() is different than from fromJson8() ? It's like @()
|
| 484 |
|
| 485 | -->
|
| 486 |
|
| 487 | <!--
|
| 488 |
|
| 489 | What about write? These would be the same:
|
| 490 |
|
| 491 | write --json -- $s
|
| 492 | write --j8 -- $s
|
| 493 |
|
| 494 | write -- $[toJson(s)]
|
| 495 | write -- $[toJson8(s)]
|
| 496 |
|
| 497 | write --json -- @strs
|
| 498 | write --j8 -- @strs
|
| 499 |
|
| 500 | write -- @[toJson(s) for s in strs]
|
| 501 | write -- @[toJson8(s) for s in strs]
|
| 502 |
|
| 503 | It's an argument for getting rid --json and --j8? I already implemented them,
|
| 504 | but it makes the API smaller.
|
| 505 |
|
| 506 | I guess the main thing would be to AVOID quoting sometimes?
|
| 507 |
|
| 508 | $ write --j8 -- unquoted
|
| 509 | unquoted
|
| 510 |
|
| 511 | $ write --j8 -- $'\'' '"'
|
| 512 | "'"
|
| 513 | "\""
|
| 514 |
|
| 515 | I think this could be the shell style?
|
| 516 |
|
| 517 | $ write --shell-str -- foo bar baz
|
| 518 |
|
| 519 | Or it could be
|
| 520 |
|
| 521 | $ write -- @[toShellString(s) for s in strs]
|
| 522 |
|
| 523 | I want this to be "J8 Lines", but it can be done in pure YSH. It's not built
|
| 524 | into the interpreter.
|
| 525 |
|
| 526 | foo/bar
|
| 527 | "hi"
|
| 528 | b'hi'
|
| 529 | u'hi'
|
| 530 |
|
| 531 | But what about
|
| 532 |
|
| 533 | Fool's Gold
|
| 534 | a'hi' # This feels like an error?
|
| 535 | a"hi" # what about this?
|
| 536 |
|
| 537 | Technically we CAN read those as literal strings
|
| 538 | -->
|
| 539 |
|
| 540 | ### ysh-echo
|
| 541 |
|
| 542 | Print arguments to stdout, separated by a space.
|
| 543 |
|
| 544 | ysh$ echo hi there
|
| 545 | hi there
|
| 546 |
|
| 547 | The [simple_echo][] option means that flags aren't accepted, and `--` is not
|
| 548 | accepted.
|
| 549 |
|
| 550 | ysh$ echo -n
|
| 551 | -n
|
| 552 |
|
| 553 | See the [YSH FAQ entry on echo][echo-en] for details.
|
| 554 |
|
| 555 | [simple_echo]: chap-option.html#ysh:all
|
| 556 | [echo-en]: ../ysh-faq.html#how-do-i-write-the-equivalent-of-echo-e-or-echo-n
|
| 557 |
|
| 558 | ### ysh-test
|
| 559 |
|
| 560 | The YSH [test](#test) builtin supports these long flags:
|
| 561 |
|
| 562 | --dir same as -d
|
| 563 | --exists same as -e
|
| 564 | --file same as -f
|
| 565 | --symlink same as -L
|
| 566 |
|
| 567 | --true Is the argument equal to the string "true"?
|
| 568 | --false Is the argument equal to the string "false"?
|
| 569 |
|
| 570 | The `--true` and `--false` flags can be used to combine commands and
|
| 571 | expressions:
|
| 572 |
|
| 573 | if test --file a && test --true $[bool(mydict)] {
|
| 574 | echo ok
|
| 575 | }
|
| 576 |
|
| 577 | This works because the boolean `true` *stringifies* to `"true"`, and likewise
|
| 578 | with `false`.
|
| 579 |
|
| 580 | That is, `$[true] === "true"` and `$[false] === "false"`.
|
| 581 |
|
| 582 | ### write
|
| 583 |
|
| 584 | write fixes problems with shell's `echo` builtin.
|
| 585 |
|
| 586 | The default separator is a newline, and the default terminator is a
|
| 587 | newline.
|
| 588 |
|
| 589 | Examples:
|
| 590 |
|
| 591 | write -- ale bean # write two lines
|
| 592 |
|
| 593 | write -n -- ale bean # synonym for --end '', like echo -n
|
| 594 | write --sep '' --end '' -- a b # write 2 bytes
|
| 595 | write --sep $'\t' --end $'\n' -- a b # TSV line
|
| 596 |
|
| 597 | You may want to use `toJson8()` or `toJson()` before writing:
|
| 598 |
|
| 599 | write -- $[toJson8(mystr)]
|
| 600 | write -- $[toJson(mystr)]
|
| 601 |
|
| 602 |
|
| 603 | <!--
|
| 604 | write --json -- ale bean # JSON encode, guarantees two lines
|
| 605 | write --j8 -- ale bean # J8 encode, guarantees two lines
|
| 606 | -->
|
| 607 |
|
| 608 |
|
| 609 | ### fork
|
| 610 |
|
| 611 | Run a command, but don't wait for it to finish.
|
| 612 |
|
| 613 | fork { sleep 1 }
|
| 614 | wait -n
|
| 615 |
|
| 616 | In YSH, use `fork` rather than shell's `&` ([ampersand][]).
|
| 617 |
|
| 618 | [ampersand]: chap-cmd-lang.html#ampersand
|
| 619 |
|
| 620 | ### forkwait
|
| 621 |
|
| 622 | The preferred alternative to shell's `()`. Prefer `cd` with a block if possible.
|
| 623 |
|
| 624 | forkwait {
|
| 625 | not_mutated=zzz
|
| 626 | }
|
| 627 | echo $not_mutated
|
| 628 |
|
| 629 | ### redir
|
| 630 |
|
| 631 | Runs a block passed to it. It's designed to enable a **prefix** syntax when
|
| 632 | redirecting:
|
| 633 |
|
| 634 | redir >out.txt {
|
| 635 | echo 1
|
| 636 | echo 2
|
| 637 | }
|
| 638 |
|
| 639 | When a block is long, it's more readable than shell's postfix style:
|
| 640 |
|
| 641 | { echo 1
|
| 642 | echo 2
|
| 643 | } >out.txt
|
| 644 |
|
| 645 | ## Private
|
| 646 |
|
| 647 | Private builtins are not enabled by default:
|
| 648 |
|
| 649 | sleep 0.1 # runs external process; private builtin not found
|
| 650 | builtin sleep 0.1 # runs private builtin
|
| 651 |
|
| 652 | ### cat
|
| 653 |
|
| 654 | `cat` is a *private* builtin that reads from files and writes to stdout.
|
| 655 |
|
| 656 | cat FILE+ # Read from each file, and write to stdout
|
| 657 | # If the file is -, read from stdin (not the file called -)
|
| 658 | cat # equivalent to cat -
|
| 659 |
|
| 660 | - Related: [rewrite_extern][]
|
| 661 |
|
| 662 | [rewrite_extern]: chap-option.html#rewrite_extern
|
| 663 |
|
| 664 | ### rm
|
| 665 |
|
| 666 | `rm` is a *private* builtin that removes files.
|
| 667 |
|
| 668 | rm FLAG* FILE*
|
| 669 |
|
| 670 | Flags:
|
| 671 |
|
| 672 | -f Don't fail if the file exist, and don't fail if no arguments are
|
| 673 | passed.
|
| 674 |
|
| 675 | Return 0 on success, and non-zero otherwise.
|
| 676 |
|
| 677 | - Related: [rewrite_extern][]
|
| 678 |
|
| 679 | ### sleep
|
| 680 |
|
| 681 | `sleep` is a *private* builtin that puts the shell process to sleep for the
|
| 682 | given number of seconds.
|
| 683 |
|
| 684 | Example:
|
| 685 |
|
| 686 | builtin sleep 0.1 # wait 100 milliseconds
|
| 687 |
|
| 688 | It respects signals:
|
| 689 |
|
| 690 | - `SIGINT` / Ctrl-C cancels the command, with the standard behavior:
|
| 691 | - in an interactive shell, you return to the prompt
|
| 692 | - a non-interactive shell is cancelled
|
| 693 | - Upon receiving other signals, Oils run pending traps, and then continues to
|
| 694 | sleep.
|
| 695 |
|
| 696 | It's compatible with the POSIX `sleep` utility:
|
| 697 |
|
| 698 | sleep 2 # wait 2 seconds
|
| 699 |
|
| 700 | ## Hay Config
|
| 701 |
|
| 702 | ### hay
|
| 703 |
|
| 704 | ### haynode
|
| 705 |
|
| 706 |
|
| 707 | ## Data Formats
|
| 708 |
|
| 709 | ### json
|
| 710 |
|
| 711 | Write JSON:
|
| 712 |
|
| 713 | var d = {name: 'bob', age: 42}
|
| 714 | json write (d) # default indent of 2, type errors
|
| 715 | json write (d, space=0) # no indent
|
| 716 | json write (d, type_errors=false) # non-serializable types become null
|
| 717 | # (e.g. Obj, Proc, Eggex)
|
| 718 |
|
| 719 | Read JSON:
|
| 720 |
|
| 721 | echo hi | json read # fills $_reply by default
|
| 722 |
|
| 723 | Or use an explicit place:
|
| 724 |
|
| 725 | var x = ''
|
| 726 | json read (&x) < myfile.txt
|
| 727 |
|
| 728 | Related: [err-json-encode][] and [err-json-decode][]
|
| 729 |
|
| 730 | [err-json-encode]: chap-errors.html#err-json-encode
|
| 731 | [err-json-decode]: chap-errors.html#err-json-decode
|
| 732 |
|
| 733 | ### json8
|
| 734 |
|
| 735 | Like `json`, but on the encoding side:
|
| 736 |
|
| 737 | - Falls back to `b'\yff'` instead of lossy Unicode replacement char
|
| 738 |
|
| 739 | On decoding side:
|
| 740 |
|
| 741 | - Understands `b'' u''` strings
|
| 742 |
|
| 743 | Related: [err-json8-encode]() and [err-json8-decode]()
|
| 744 |
|
| 745 | [err-json8-encode]: chap-errors.html#err-json8-encode
|
| 746 | [err-json8-decode]: chap-errors.html#err-json8-decode
|
| 747 |
|
| 748 | ## I/O
|
| 749 |
|
| 750 | These builtins take input and output. They're often used with redirects.
|
| 751 |
|
| 752 | ### read
|
| 753 |
|
| 754 | read FLAG* VAR*
|
| 755 |
|
| 756 | Read input from `stdin`, and assign pieces of input to variables. Without
|
| 757 | flags, `read` uses this algorithm:
|
| 758 |
|
| 759 | 1. Read bytes from `stdin`, one at a time, until a newline `\n`.
|
| 760 | - Respect `\` escapes and line continuations.
|
| 761 | - Any NUL bytes are removed from the input.
|
| 762 | 1. Use the `$IFS` algorithm to split the line into N pieces, where `N` is the
|
| 763 | number of `VAR` specified. Each piece is assigned to the corresponding
|
| 764 | variable.
|
| 765 | - If no VARs are given, assign to the `$REPLY` var.
|
| 766 |
|
| 767 | The `-r` flag is useful to disable backslash escapes.
|
| 768 |
|
| 769 | POSIX mandates the slow behavior of reading one byte at a time. In YSH, you
|
| 770 | can avoid this by using [io.stdin][], or a `--long-flag` documented in
|
| 771 | [ysh-read](#ysh-read).
|
| 772 |
|
| 773 | Flags:
|
| 774 |
|
| 775 | -a ARRAY assign the tokens to elements of this array
|
| 776 | -d CHAR use DELIM as delimiter, instead of newline
|
| 777 | -n NUM read up to NUM characters, respecting delimiters. When -r is not
|
| 778 | specified, backslash escape of the form "\?" is counted as one
|
| 779 | character. This is the Bash behavior, but other shells such as
|
| 780 | ash and mksh count the number of bytes with "-n" without
|
| 781 | considering backslash escaping.
|
| 782 | -p STR print the string PROMPT before reading input
|
| 783 | -r raw mode: don't let backslashes escape characters
|
| 784 | -s silent: do not echo input coming from a terminal
|
| 785 | -t NUM time out and fail after TIME seconds
|
| 786 | -t 0 returns whether any input is available
|
| 787 | -u FD read from file descriptor FD instead of 0 (stdin)
|
| 788 |
|
| 789 | <!-- -N NUM read up to NUM characters, ignoring delimiters -->
|
| 790 | <!-- -e use readline to obtain the line
|
| 791 | -i STR use STR as the initial text for readline -->
|
| 792 |
|
| 793 | Performance summary: [YSH Input/Output > Three Types of I/O][ysh-io-three]
|
| 794 |
|
| 795 | ### echo
|
| 796 |
|
| 797 | echo FLAG* ARG*
|
| 798 |
|
| 799 | Prints ARGs to stdout, separated by a space, and terminated by a newline.
|
| 800 |
|
| 801 | Flags:
|
| 802 |
|
| 803 | -e enable interpretation of backslash escapes
|
| 804 | -n omit the trailing newline
|
| 805 | <!-- -E -->
|
| 806 |
|
| 807 | `echo` in YSH does **not** accept these flags. See [ysh-echo](#ysh-echo) and
|
| 808 | [the FAQ entry][echo-en]. (This is unusual because YSH doesn't usually "break"
|
| 809 | OSH.)
|
| 810 |
|
| 811 | See [char-escapes](chap-mini-lang.html#char-escapes) to see what's supported
|
| 812 | when `-e` is passed.
|
| 813 |
|
| 814 | ### printf
|
| 815 |
|
| 816 | printf FLAG* FMT ARG*
|
| 817 |
|
| 818 | Formats values and prints them. The FMT string contain three types of objects:
|
| 819 |
|
| 820 | 1. Literal Characters
|
| 821 | 2. Character escapes like `\t`. See [char-escapes](chap-mini-lang.html#char-escapes).
|
| 822 | 3. Percent codes like `%s` that specify how to format each each ARG.
|
| 823 |
|
| 824 | If not enough ARGS are passed, the empty string is used. If too many are
|
| 825 | passed, the FMT string will be "recycled".
|
| 826 |
|
| 827 | Flags:
|
| 828 |
|
| 829 | -v VAR Write output in variable VAR instead of standard output.
|
| 830 |
|
| 831 | Format specifiers:
|
| 832 |
|
| 833 | %% Prints a single "%".
|
| 834 | %b Interprets backslash escapes while printing.
|
| 835 | %q Prints the argument escaping the characters needed to make it reusable
|
| 836 | as shell input.
|
| 837 | %d Print as signed decimal number.
|
| 838 | %i Same as %d.
|
| 839 | %o Print as unsigned octal number.
|
| 840 | %u Print as unsigned decimal number.
|
| 841 | %x Print as unsigned hexadecimal number with lower-case hex-digits (a-f).
|
| 842 | %X Same as %x, but with upper-case hex-digits (A-F).
|
| 843 | %f Print as floating point number.
|
| 844 | %e Print as a double number, in "±e" format (lower-case e).
|
| 845 | %E Same as %e, but with an upper-case E.
|
| 846 | %g Interprets the argument as double, but prints it like %f or %e.
|
| 847 | %G Same as %g, but print it like %E.
|
| 848 | %c Print as a single char, only the first character is printed.
|
| 849 | %s Print as string
|
| 850 | %n The number of characters printed so far is stored in the variable named
|
| 851 | in the argument.
|
| 852 | %a Interprets the argument as double, and prints it like a C99 hexadecimal
|
| 853 | floating-point literal.
|
| 854 | %A Same as %a, but print it like %E.
|
| 855 | %(FORMAT)T Prints date and time, according to FORMAT as a format string
|
| 856 | for strftime(3). The argument is the number of seconds since
|
| 857 | epoch. It can also be -1 (current time, also the default value
|
| 858 | if there is no argument) or -2 (shell startup time).
|
| 859 |
|
| 860 | ### readarray
|
| 861 |
|
| 862 | Alias for `mapfile`.
|
| 863 |
|
| 864 | ### mapfile
|
| 865 |
|
| 866 | mapfile FLAG* ARRAY?
|
| 867 |
|
| 868 | Reads lines from stdin into the variable named ARRAY (default
|
| 869 | `${MAPFILE[@]}`).
|
| 870 |
|
| 871 | Flags:
|
| 872 |
|
| 873 | -t Remove the trailing newline from every line
|
| 874 | <!--
|
| 875 | -d CHAR use CHAR as delimiter, instead of the default newline
|
| 876 | -n NUM copy up to NUM lines
|
| 877 | -O NUM begins copying lines at the NUM element of the array
|
| 878 | -s NUM discard the first NUM lines
|
| 879 | -u FD read from FD file descriptor instead of the standard input
|
| 880 | -C CMD run CMD every NUM lines specified in -c
|
| 881 | -c NUM every NUM lines, the CMD command in C will be run
|
| 882 | -->
|
| 883 |
|
| 884 | ## Run Code
|
| 885 |
|
| 886 | These builtins accept shell code and run it.
|
| 887 |
|
| 888 | ### source
|
| 889 |
|
| 890 | source SCRIPT ARG*
|
| 891 |
|
| 892 | Execute SCRIPT with the given ARGs, in the context of the current shell. That is,
|
| 893 | existing variables will be modified.
|
| 894 |
|
| 895 | ---
|
| 896 |
|
| 897 | Oils extension: If the SCRIPT starts with `///`, we look for scripts embedded in
|
| 898 | the `oils-for-unix` binary. Example:
|
| 899 |
|
| 900 | source ///osh/two.sh # load embedded script
|
| 901 |
|
| 902 | : ${LIB_OSH=fallback/dir}
|
| 903 | source $LIB_OSH/two.sh # same thing
|
| 904 |
|
| 905 | The [LIB_OSH][] form is useful for writing a script that works under both bash
|
| 906 | and OSH.
|
| 907 |
|
| 908 | - Related: the [cat-em][] tool prints embedded scripts.
|
| 909 |
|
| 910 | [LIB_OSH]: chap-special-var.html#LIB_OSH
|
| 911 | [cat-em]: chap-front-end.html#cat-em
|
| 912 |
|
| 913 |
|
| 914 | ### cmd/eval
|
| 915 |
|
| 916 | eval ARG+
|
| 917 |
|
| 918 | Creates a string by joining ARGs with a space, then runs it as a shell command.
|
| 919 |
|
| 920 | Example:
|
| 921 |
|
| 922 | # Create the string echo "hello $name" and run it.
|
| 923 | a='echo'
|
| 924 | b='"hello $name"'
|
| 925 | eval $a $b
|
| 926 |
|
| 927 | Tips:
|
| 928 |
|
| 929 | - Using `eval` can confuse code and user-supplied data, leading to [security
|
| 930 | issues][].
|
| 931 | - Prefer passing single string ARG to `eval`.
|
| 932 |
|
| 933 | [security issues]: https://mywiki.wooledge.org/BashFAQ/048
|
| 934 |
|
| 935 | ### trap
|
| 936 |
|
| 937 | The `trap` builtin lets you run shell code when events happen. Events are
|
| 938 | signals or interpreter hooks.
|
| 939 |
|
| 940 | trap -l # List all events and their number
|
| 941 | trap -p # Print the current trap state: events and handlers
|
| 942 | trap CMD EVENT* # Register handlers
|
| 943 | trap - EVENT* # Remove handlers
|
| 944 |
|
| 945 | Examples:
|
| 946 |
|
| 947 | trap 'echo hi' EXIT INT # Register
|
| 948 | trap - EXIT INT # Remove
|
| 949 |
|
| 950 | OSH also support legacy syntax, which is not recommended:
|
| 951 |
|
| 952 | trap 'echo hi' 0 # 0 is the exit trap
|
| 953 | trap INT # remove signal handler
|
| 954 | trap 0 # remove exit trap
|
| 955 | trap 0 INT # remove both
|
| 956 |
|
| 957 | <!--
|
| 958 | trap '' EVENT* # TODO Ignore events
|
| 959 | -->
|
| 960 |
|
| 961 | Tips:
|
| 962 |
|
| 963 | - Prefer passing the name of a shell function to `trap`.
|
| 964 | - See [ysh-trap](#ysh-trap) for even nicer idioms.
|
| 965 | - See [Chapter: Plugins and Hooks > Traps](chap-plugin.html#Traps) for a list of
|
| 966 | traps, like `trap '' EXIT`.
|
| 967 |
|
| 968 | ### ysh-trap
|
| 969 |
|
| 970 | The `trap` builtin lets you run shell code when events happen.
|
| 971 |
|
| 972 | YSH improves the syntax of the trap builtin, and removes legacy.
|
| 973 |
|
| 974 | trap -l # List all events and their number
|
| 975 | trap -p # Print the current trap state: events and handlers
|
| 976 | trap --add EVENT* BLOCK # Register handlers
|
| 977 | trap --remove EVENT* # Remove handlers
|
| 978 |
|
| 979 | Examples:
|
| 980 |
|
| 981 | trap --add EXIT INT {
|
| 982 | echo 'either exit'
|
| 983 | echo 'or int'
|
| 984 | }
|
| 985 |
|
| 986 | trap --remove EXIT INT
|
| 987 |
|
| 988 | Note: the block argument to `trap --add` doesn't capture variables -- it's not
|
| 989 | a closure. So YSH behaves like OSH, but the syntax doesn't encourage putting
|
| 990 | source code in strings.
|
| 991 |
|
| 992 | ## Set Options
|
| 993 |
|
| 994 | The `set` and `shopt` builtins set global shell options. YSH code should use
|
| 995 | the more natural `shopt`.
|
| 996 |
|
| 997 | ### set
|
| 998 |
|
| 999 | set FLAG* ARG*
|
| 1000 |
|
| 1001 | Sets global shell options. Short style:
|
| 1002 |
|
| 1003 | set -e
|
| 1004 |
|
| 1005 | Long style:
|
| 1006 |
|
| 1007 | set -o errexit
|
| 1008 |
|
| 1009 | Set the arguments array:
|
| 1010 |
|
| 1011 | set -- 1 2 3
|
| 1012 |
|
| 1013 | See [Chapter: Global Shell Options](chap-option.html) for a list of options.
|
| 1014 |
|
| 1015 | ### shopt
|
| 1016 |
|
| 1017 | shopt FLAG* OPTION* BLOCK?
|
| 1018 |
|
| 1019 | Sets global shell options.
|
| 1020 |
|
| 1021 | Flags:
|
| 1022 |
|
| 1023 | -s --set Turn the named options on
|
| 1024 | -u --unset Turn the named options off
|
| 1025 | -p Print option values, and 1 if any option is unset
|
| 1026 | -o Use older set of options, normally controlled by 'set -o'
|
| 1027 | -q Return 0 if the option is true, else 1
|
| 1028 |
|
| 1029 | This command is compatible with `shopt` in bash. See [ysh-shopt](#ysh-shopt) for
|
| 1030 | details on YSH enhancements.
|
| 1031 |
|
| 1032 | See [Chapter: Global Shell Options](chap-option.html) for a list of options.
|
| 1033 |
|
| 1034 | ## Working Dir
|
| 1035 |
|
| 1036 | These 5 builtins deal with the working directory of the shell.
|
| 1037 |
|
| 1038 | ### cd
|
| 1039 |
|
| 1040 | cd FLAG* DIR
|
| 1041 |
|
| 1042 | Changes the working directory of the current shell process to DIR.
|
| 1043 |
|
| 1044 | If DIR isn't specified, change to `$HOME`. If DIR is `-`, change to `$OLDPWD`
|
| 1045 | (a variable that the sets to the previous working directory.)
|
| 1046 |
|
| 1047 | Flags:
|
| 1048 |
|
| 1049 | -L Follow symbolic links, i.e. change to the TARGET of the symlink.
|
| 1050 | (default).
|
| 1051 | -P Don't follow symbolic links.
|
| 1052 |
|
| 1053 | ### chdir
|
| 1054 |
|
| 1055 | `chdir` is a synonym for `cd`. Shells like `busybox ash` support it, so OSH
|
| 1056 | does too.
|
| 1057 |
|
| 1058 | ### pwd
|
| 1059 |
|
| 1060 | pwd FLAG*
|
| 1061 |
|
| 1062 | Prints the current working directory.
|
| 1063 |
|
| 1064 | Flags:
|
| 1065 |
|
| 1066 | -L Follow symbolic links if present (default)
|
| 1067 | -P Don't follow symbolic links. Print the link instead of the target.
|
| 1068 |
|
| 1069 | ### pushd
|
| 1070 |
|
| 1071 | <!--pushd FLAGS DIR-->
|
| 1072 | pushd DIR
|
| 1073 | <!--pushd +/-NUM-->
|
| 1074 |
|
| 1075 | Add DIR to the directory stack, then change the working directory to DIR.
|
| 1076 | Typically used with `popd` and `dirs`.
|
| 1077 |
|
| 1078 | <!--FLAGS:
|
| 1079 | -n Don't change the working directory, just manipulate the stack
|
| 1080 | NUM:
|
| 1081 | Rotates the stack the number of places specified. Eg, given the stack
|
| 1082 | '/foo /bar /baz', where '/foo' is the top of the stack, pushd +1 will move
|
| 1083 | it to the bottom, '/bar /baz /foo'-->
|
| 1084 |
|
| 1085 | ### popd
|
| 1086 |
|
| 1087 | popd
|
| 1088 |
|
| 1089 | Removes a directory from the directory stack, and changes the working directory
|
| 1090 | to it. Typically used with `pushd` and `dirs`.
|
| 1091 |
|
| 1092 | ### dirs
|
| 1093 |
|
| 1094 | dirs FLAG*
|
| 1095 |
|
| 1096 | Shows the contents of the directory stack. Typically used with `pushd` and
|
| 1097 | `popd`.
|
| 1098 |
|
| 1099 | Flags:
|
| 1100 |
|
| 1101 | -c Clear the dir stack.
|
| 1102 | -l Show the dir stack, but with the real path instead of ~.
|
| 1103 | -p Show the dir stack, but formatted as one line per entry.
|
| 1104 | -v Like -p, but numbering each line.
|
| 1105 |
|
| 1106 | ## Completion
|
| 1107 |
|
| 1108 | These builtins implement our bash-compatible autocompletion system.
|
| 1109 |
|
| 1110 | ### complete
|
| 1111 |
|
| 1112 | Registers completion policies for different commands.
|
| 1113 |
|
| 1114 | ### compgen
|
| 1115 |
|
| 1116 | Generates completion candidates inside a user-defined completion function.
|
| 1117 |
|
| 1118 | It can also be used in scripts, i.e. outside a completion function.
|
| 1119 |
|
| 1120 | ### compopt
|
| 1121 |
|
| 1122 | Changes completion options inside a user-defined completion function.
|
| 1123 |
|
| 1124 | ### compadjust
|
| 1125 |
|
| 1126 | Adjusts `COMP_ARGV` according to specified delimiters, and optionally set
|
| 1127 | variables cur, prev, words (an array), and cword. May also set 'split'.
|
| 1128 |
|
| 1129 | This is an OSH extension that makes it easier to run the bash-completion
|
| 1130 | project.
|
| 1131 |
|
| 1132 | ### compexport
|
| 1133 |
|
| 1134 | Complete an entire shell command string. For example,
|
| 1135 |
|
| 1136 | compexport -c 'echo $H'
|
| 1137 |
|
| 1138 | will complete variables like `$HOME`. And
|
| 1139 |
|
| 1140 | compexport -c 'ha'
|
| 1141 |
|
| 1142 | will complete builtins like `hay`, as well as external commands.
|
| 1143 |
|
| 1144 |
|
| 1145 | ## Shell Process
|
| 1146 |
|
| 1147 | These builtins mutate the state of the shell process.
|
| 1148 |
|
| 1149 | ### exec
|
| 1150 |
|
| 1151 | exec BIN_PATH ARG*
|
| 1152 |
|
| 1153 | Replaces the running shell with the binary specified, which is passed ARGs.
|
| 1154 | BIN_PATH must exist on the file system; i.e. it can't be a shell builtin or
|
| 1155 | function.
|
| 1156 |
|
| 1157 | ### umask
|
| 1158 |
|
| 1159 | umask OCTAL
|
| 1160 | umask SYMBOLIC[,SYMBOLIC]...
|
| 1161 |
|
| 1162 | umask FLAG* # print the mask
|
| 1163 |
|
| 1164 | Sets the bit mask that determines the permissions for new files and
|
| 1165 | directories. The mask is subtracted from 666 for files and 777 for
|
| 1166 | directories.
|
| 1167 |
|
| 1168 | If no MODE, show the current mask.
|
| 1169 |
|
| 1170 | Flags:
|
| 1171 |
|
| 1172 | -p print in a form that may be reused as input
|
| 1173 | -S print in symbolic form
|
| 1174 |
|
| 1175 | Symbolic mask:
|
| 1176 |
|
| 1177 | A SYMBOLIC clause is of the form '[ugoa]*([-+=]([rwxXstugo]*))*', the same as is accepted by chmod.
|
| 1178 |
|
| 1179 | who :: u | g | o | a
|
| 1180 | operator :: + | - | =
|
| 1181 | permission :: r | w | x | X | s | t | u | g | o
|
| 1182 |
|
| 1183 | Who:
|
| 1184 | ugo set the user, group, or other parts of the file mode bits, respectively
|
| 1185 | a equivalent to all of ugo
|
| 1186 |
|
| 1187 | Operator:
|
| 1188 | + removes specified permission bits from the umask according to who
|
| 1189 | - adds specified permission bits to the umask according to who
|
| 1190 | = set all who permissions to 0777, then remove specified permission bits from the umask according to who
|
| 1191 |
|
| 1192 | Permission:
|
| 1193 | rwx set the read, write, or execute bit respectively
|
| 1194 | X set the execute bit iff the umask before this operation has one of its 3 execute bits set
|
| 1195 | st does nothing, but exists for POSIX compatibility
|
| 1196 | ugo set permission equal to the umask's u, g, and o bits before this operation
|
| 1197 |
|
| 1198 | Examples:
|
| 1199 |
|
| 1200 | umask # get the current umask in octal form
|
| 1201 |
|
| 1202 | umask 0124
|
| 1203 |
|
| 1204 | umask ug=rwx,o-r # clauses may be joined using commas
|
| 1205 |
|
| 1206 | umask u+rwx-x-w-r # operators can be chained
|
| 1207 |
|
| 1208 | umask a=u # [ugo] after the operator represents the initial umask permissions
|
| 1209 |
|
| 1210 | umask =u # equivalent to `umask a=u`
|
| 1211 |
|
| 1212 | ### ulimit
|
| 1213 |
|
| 1214 | ulimit --all
|
| 1215 | ulimit -a
|
| 1216 | ulimit FLAGS* -RESOURCE_FLAG VALUE?
|
| 1217 |
|
| 1218 | ulimit FLAGS* VALUE? # discouraged
|
| 1219 |
|
| 1220 | Show and modify process resource limits.
|
| 1221 |
|
| 1222 | Flags:
|
| 1223 |
|
| 1224 | -S for soft limit
|
| 1225 | -H for hard limit
|
| 1226 |
|
| 1227 | -c -d -f ... # ulimit --all shows all resource flags
|
| 1228 |
|
| 1229 | Show a table of resources:
|
| 1230 |
|
| 1231 | ulimit --all
|
| 1232 | ulimit -a
|
| 1233 |
|
| 1234 | For example, the table shows that `-n` is the flag that controls the number
|
| 1235 | file descriptors, the soft and hard limit for `-n`, and the multiplication
|
| 1236 | "factor" for the integer VALUE you pass.
|
| 1237 |
|
| 1238 | ---
|
| 1239 |
|
| 1240 | Here are examples of using resource flags.
|
| 1241 |
|
| 1242 | Get the soft limit for the number of file descriptors:
|
| 1243 |
|
| 1244 | ulimit -S -n
|
| 1245 | ulimit -n # same thing
|
| 1246 |
|
| 1247 | Get the hard limit:
|
| 1248 |
|
| 1249 | ulimit -H -n
|
| 1250 |
|
| 1251 | Set the soft or hard limit:
|
| 1252 |
|
| 1253 | ulimit -S -n 100
|
| 1254 | ulimit -H -n 100
|
| 1255 |
|
| 1256 | Set both limits:
|
| 1257 |
|
| 1258 | ulimit -n 100
|
| 1259 |
|
| 1260 | A special case that's discouraged: with no resource flag, `-f` is assumed:
|
| 1261 |
|
| 1262 | ulimit # equivalent to ulimit -f
|
| 1263 | ulimit 100 # equivalent to ulimit -f 100
|
| 1264 |
|
| 1265 | ### times
|
| 1266 |
|
| 1267 | times
|
| 1268 |
|
| 1269 | Shows the user and system time used by the shell and its child processes.
|
| 1270 |
|
| 1271 | ## Child Process
|
| 1272 |
|
| 1273 | ### jobs
|
| 1274 |
|
| 1275 | jobs
|
| 1276 |
|
| 1277 | Shows all jobs running in the shell and their status.
|
| 1278 |
|
| 1279 | ### wait
|
| 1280 |
|
| 1281 | Wait for jobs to finish, in a few different ways. (A job is a process or a
|
| 1282 | pipeline.)
|
| 1283 |
|
| 1284 | wait # no arguments
|
| 1285 |
|
| 1286 | Wait for all jobs to terminate. The exit status is 0, unless a signal occurs.
|
| 1287 |
|
| 1288 | wait -n
|
| 1289 |
|
| 1290 | Wait for the next job to terminate, and return its status.
|
| 1291 |
|
| 1292 | wait $pid1 $pid2 ...
|
| 1293 |
|
| 1294 | Wait for the jobs specified by PIDs to terminate. Return the status of the
|
| 1295 | last one.
|
| 1296 |
|
| 1297 | wait %3 %2 ...
|
| 1298 |
|
| 1299 | Wait for the jobs specified by "job specs" to terminate. Return the status of
|
| 1300 | the last one.
|
| 1301 |
|
| 1302 | ---
|
| 1303 |
|
| 1304 | If wait is interrupted by a signal, the exit status is the signal number + 128.
|
| 1305 |
|
| 1306 | ---
|
| 1307 |
|
| 1308 | When using `set -e` aka `errexit`, `wait --all` is useful. See topic
|
| 1309 | [ysh-wait](#ysh-wait).
|
| 1310 |
|
| 1311 | <!--
|
| 1312 | The ARG can be a PID (tracked by the kernel), or a job number (tracked by the
|
| 1313 | shell). Specify jobs with the syntax `%jobnumber`.
|
| 1314 | -->
|
| 1315 |
|
| 1316 | ### ysh-wait
|
| 1317 |
|
| 1318 | YSH extends the `wait` builtin with 2 flags:
|
| 1319 |
|
| 1320 | wait --all # wait for all jobs, like 'wait'
|
| 1321 | # but exit 1 if any job exits non-zero
|
| 1322 |
|
| 1323 | wait --verbose # show a message on each job completion
|
| 1324 |
|
| 1325 | wait --all --verbose # show a message, and also respect failure
|
| 1326 |
|
| 1327 | ### fg
|
| 1328 |
|
| 1329 | fg JOB?
|
| 1330 |
|
| 1331 | Continues a stopped job in the foreground. This means it can receive signals
|
| 1332 | from the keyboard, like Ctrl-C and Ctrl-Z.
|
| 1333 |
|
| 1334 | If no JOB is specified, use the latest job.
|
| 1335 |
|
| 1336 | ### bg
|
| 1337 |
|
| 1338 | UNIMPLEMENTED
|
| 1339 |
|
| 1340 | bg JOB?
|
| 1341 |
|
| 1342 | Continues a stopped job, while keeping it in the background. This means it
|
| 1343 | **can't** receive signals from the keyboard, like Ctrl-C and Ctrl-Z.
|
| 1344 |
|
| 1345 | If no JOB is specified, use the latest job.
|
| 1346 |
|
| 1347 | ### kill
|
| 1348 |
|
| 1349 | The `kill` builtin sends a signal to one or more processes. Usage:
|
| 1350 |
|
| 1351 | kill (-s SIG | -SIG)? WHAT+ # send SIG to the given processes
|
| 1352 |
|
| 1353 | where
|
| 1354 |
|
| 1355 | SIG = NAME | NUMBER # e.g. USR1 or 10
|
| 1356 | WHAT = PID | JOBSPEC # e.g. 789 or %%
|
| 1357 |
|
| 1358 | Examples:
|
| 1359 |
|
| 1360 | kill -s USR1 789 # send SIGUSR1 to PID 789
|
| 1361 |
|
| 1362 | kill -s USR1 789 %% # send signal to PID 789 and the current job
|
| 1363 | kill -s 10 789 %% # specify SIGUSR1 by number instead
|
| 1364 |
|
| 1365 | kill -USR1 789 %% # shortcut syntax
|
| 1366 | kill -10 789 %% # shortcut using a number
|
| 1367 |
|
| 1368 | kill -n USR1 789 %% # -n is a synonym for -s
|
| 1369 | kill 789 %% # if not specified, the default is SIGTERM
|
| 1370 |
|
| 1371 | ---
|
| 1372 |
|
| 1373 | It can also list signals:
|
| 1374 |
|
| 1375 | kill -L # List all signals
|
| 1376 | kill -L SIG+ # Translate signals from name to number, and vice versa
|
| 1377 |
|
| 1378 | Examples:
|
| 1379 |
|
| 1380 | kill -l # List all signals; -l is a synonym for -L
|
| 1381 | kill -L USR1 USR2 # prints '10 12'
|
| 1382 | kill -L USR1 15 # prints '10 TERM'
|
| 1383 |
|
| 1384 | ## External
|
| 1385 |
|
| 1386 | ### test
|
| 1387 |
|
| 1388 | test OP ARG
|
| 1389 | test ARG OP ARG
|
| 1390 | [ OP ARG ] # [ is an alias for test that requires closing ]
|
| 1391 | [ ARG OP ARG ]
|
| 1392 |
|
| 1393 | Evaluates a conditional expression and returns 0 (true) or 1 (false).
|
| 1394 |
|
| 1395 | Note that `[` is the name of a builtin, not an operator in the language. Use
|
| 1396 | `test` to avoid this confusion.
|
| 1397 |
|
| 1398 | String expressions:
|
| 1399 |
|
| 1400 | -n STR True if STR is not empty.
|
| 1401 | 'test STR' is usually equivalent, but discouraged.
|
| 1402 | -z STR True if STR is empty.
|
| 1403 | STR1 = STR2 True if the strings are equal.
|
| 1404 | STR1 != STR2 True if the strings are not equal.
|
| 1405 | STR1 < STR2 True if STR1 sorts before STR2 lexicographically.
|
| 1406 | STR1 > STR2 True if STR1 sorts after STR2 lexicographically.
|
| 1407 | Note: < and > should be quoted like \< and \>
|
| 1408 |
|
| 1409 | File expressions:
|
| 1410 |
|
| 1411 | -a FILE Synonym for -e.
|
| 1412 | -b FILE True if FILE is a block special file.
|
| 1413 | -c FILE True if FILE is a character special file.
|
| 1414 | -d FILE True if FILE is a directory.
|
| 1415 | -e FILE True if FILE exists.
|
| 1416 | -f FILE True if FILE is a regular file.
|
| 1417 | -g FILE True if FILE has the sgid bit set.
|
| 1418 | -G FILE True if current user's group is also FILE's group.
|
| 1419 | -h FILE True if FILE is a symbolic link.
|
| 1420 | -L FILE True if FILE is a symbolic link.
|
| 1421 | -k FILE True if FILE has the sticky bit set.
|
| 1422 | -O FILE True if current user is the file owner.
|
| 1423 | -p FILE True if FILE is a named pipe (FIFO).
|
| 1424 | -r FILE True if FILE is readable.
|
| 1425 | -s FILE True if FILE has size bigger than 0.
|
| 1426 | -S FILE True if FILE is a socket file.
|
| 1427 | -t FD True if file descriptor FD is open and refers to a terminal.
|
| 1428 | -u FILE True if FILE has suid bit set.
|
| 1429 | -w FILE True if FILE is writable.
|
| 1430 | -x FILE True if FILE is executable.
|
| 1431 | FILE1 -nt FILE2 True if FILE1 is newer than FILE2 (mtime).
|
| 1432 | FILE1 -ot FILE2 True if FILE1 is older than FILE2 (mtime).
|
| 1433 | FILE1 -ef FILE2 True if FILE1 is a hard link to FILE2.
|
| 1434 | <!-- -N FILE True if FILE was modified since last read (mtime newer than atime).-->
|
| 1435 |
|
| 1436 | Arithmetic expressions coerce arguments to integers, then compare:
|
| 1437 |
|
| 1438 | INT1 -eq INT2 True if they're equal.
|
| 1439 | INT1 -ne INT2 True if they're not equal.
|
| 1440 | INT1 -lt INT2 True if INT1 is less than INT2.
|
| 1441 | INT1 -le INT2 True if INT1 is less or equal than INT2.
|
| 1442 | INT1 -gt INT2 True if INT1 is greater than INT2.
|
| 1443 | INT1 -ge INT2 True if INT1 is greater or equal than INT2.
|
| 1444 |
|
| 1445 | Other expressions:
|
| 1446 |
|
| 1447 | -o OPTION True if the shell option OPTION is set.
|
| 1448 | -v VAR True if the variable VAR is set.
|
| 1449 |
|
| 1450 | The test builtin also supports POSIX conditionals like -a, -o, !, and ( ), but
|
| 1451 | these are discouraged.
|
| 1452 |
|
| 1453 | <!-- -R VAR True if the variable VAR has been set and is a nameref variable. -->
|
| 1454 |
|
| 1455 | ---
|
| 1456 |
|
| 1457 | See [ysh-test](#ysh-test) for log flags like `--file` and `--true`.
|
| 1458 |
|
| 1459 | ### getopts
|
| 1460 |
|
| 1461 | getopts SPEC VAR ARG*
|
| 1462 |
|
| 1463 | A single iteration of flag parsing. The SPEC is a sequence of flag characters,
|
| 1464 | with a trailing `:` to indicate that the flag takes an argument:
|
| 1465 |
|
| 1466 | ab # accept -a and -b
|
| 1467 | xy:z # accept -x, -y arg, and -z
|
| 1468 |
|
| 1469 | The input is `"$@"` by default, unless ARGs are passed.
|
| 1470 |
|
| 1471 | On each iteration, the flag character is stored in VAR. If the flag has an
|
| 1472 | argument, it's stored in `$OPTARG`. When an error occurs, VAR is set to `?`
|
| 1473 | and `$OPTARG` is unset.
|
| 1474 |
|
| 1475 | Returns 0 if a flag is parsed, or 1 on end of input or another error.
|
| 1476 |
|
| 1477 | Example:
|
| 1478 |
|
| 1479 | while getopts "ab:" flag; do
|
| 1480 | case $flag in
|
| 1481 | a) flag_a=1 ;;
|
| 1482 | b) flag_b=$OPTARG" ;;
|
| 1483 | '?') echo 'Invalid Syntax'; break ;;
|
| 1484 | esac
|
| 1485 | done
|
| 1486 |
|
| 1487 | Notes:
|
| 1488 | - `$OPTIND` is initialized to 1 every time a shell starts, and is used to
|
| 1489 | maintain state between invocations of `getopts`.
|
| 1490 | - The characters `:` and `?` can't be flags.
|
| 1491 |
|
| 1492 |
|
| 1493 | ## Conditional
|
| 1494 |
|
| 1495 | ### cmd/true
|
| 1496 |
|
| 1497 | Do nothing and return status 0.
|
| 1498 |
|
| 1499 | if true; then
|
| 1500 | echo hello
|
| 1501 | fi
|
| 1502 |
|
| 1503 | ### cmd/false
|
| 1504 |
|
| 1505 | Do nothing and return status 1.
|
| 1506 |
|
| 1507 | if false; then
|
| 1508 | echo 'not reached'
|
| 1509 | else
|
| 1510 | echo hello
|
| 1511 | fi
|
| 1512 |
|
| 1513 | <h3 id="colon" class="osh-topic">colon :</h3>
|
| 1514 |
|
| 1515 | Like `true`: do nothing and return status 0.
|
| 1516 |
|
| 1517 | ## Introspection
|
| 1518 |
|
| 1519 | <h3 id="help" class="osh-topic ysh-topic" oils-embed="1">
|
| 1520 | help
|
| 1521 | </h3>
|
| 1522 |
|
| 1523 | <!-- pre-formatted for help builtin -->
|
| 1524 |
|
| 1525 | ```
|
| 1526 | Usage: help TOPIC?
|
| 1527 |
|
| 1528 | Examples:
|
| 1529 |
|
| 1530 | help # this help
|
| 1531 | help echo # help on the 'echo' builtin
|
| 1532 | help command-sub # help on command sub $(date)
|
| 1533 |
|
| 1534 | help oils-usage # identical to oils-for-unix --help
|
| 1535 | help osh-usage # osh --help
|
| 1536 | help ysh-usage # ysh --help
|
| 1537 | ```
|
| 1538 |
|
| 1539 | ### hash
|
| 1540 |
|
| 1541 | hash
|
| 1542 |
|
| 1543 | Display information about remembered commands.
|
| 1544 |
|
| 1545 | hash FLAG* CMD+
|
| 1546 |
|
| 1547 | Determine the locations of commands using `$PATH`, and remember them.
|
| 1548 |
|
| 1549 | Flag:
|
| 1550 |
|
| 1551 | -r Discard all remembered locations.
|
| 1552 | <!-- -d Discard the remembered location of each NAME.
|
| 1553 | -l Display output in a format reusable as input.
|
| 1554 | -p PATH Inhibit path search, PATH is used as location for NAME.
|
| 1555 | -t Print the full path of one or more NAME.-->
|
| 1556 |
|
| 1557 | ### cmd/type
|
| 1558 |
|
| 1559 | type FLAG* NAME+
|
| 1560 |
|
| 1561 | Print the type of each NAME, if it were the first word of a command. Is it a
|
| 1562 | shell keyword, builtin command, shell function, alias, or executable file on
|
| 1563 | $PATH?
|
| 1564 |
|
| 1565 | Flags:
|
| 1566 |
|
| 1567 | -a Show all possible candidates, not just the first one
|
| 1568 | -f Don't search for shell functions
|
| 1569 | -P Only search for executable files
|
| 1570 | -t Print a single word: alias, builtin, file, function, proc, keyword
|
| 1571 |
|
| 1572 | Note: [`invoke --show`][invoke] is more general than `type`.
|
| 1573 |
|
| 1574 | Similar names: [type][]
|
| 1575 |
|
| 1576 | [type]: chap-index.html#type
|
| 1577 |
|
| 1578 | <!-- TODO:
|
| 1579 | - procs are counted as shell functions, should be their own thing
|
| 1580 | - Hay nodes ('hay define x') also live in the first word namespace, and should
|
| 1581 | be recognized
|
| 1582 | -->
|
| 1583 |
|
| 1584 | Modeled after the [bash `type`
|
| 1585 | builtin](https://www.gnu.org/software/bash/manual/bash.html#index-type).
|
| 1586 |
|
| 1587 | ## Word Lookup
|
| 1588 |
|
| 1589 | ### invoke
|
| 1590 |
|
| 1591 | The `invoke` builtin allows more control over name lookup than [simple
|
| 1592 | commands][simple-command].
|
| 1593 |
|
| 1594 | [simple-command]: chap-cmd-lang.html#simple-command
|
| 1595 |
|
| 1596 | Usage:
|
| 1597 |
|
| 1598 | invoke --show NAME* # Show info about EACH name
|
| 1599 | invoke NAMESPACE_FLAG+ ARG* # Run a single command with this arg array
|
| 1600 |
|
| 1601 | Namespace flags:
|
| 1602 |
|
| 1603 | --proc Run YSH procs
|
| 1604 | including invokable obj
|
| 1605 | --sh-func Run shell functions
|
| 1606 | --builtin Run builtin commands (of any kind)
|
| 1607 | eval : POSIX special
|
| 1608 | cd : normal
|
| 1609 | sleep: private (Oils)
|
| 1610 | --extern Run external commands, like /bin/ls
|
| 1611 |
|
| 1612 | Multiple namespace flags may be passed. They are searched in that order:
|
| 1613 | procs, shell functions, builtins, then extern. The first one wins. (This is
|
| 1614 | different than [command-lookup-order][].)
|
| 1615 |
|
| 1616 | [command-lookup-order]: chap-cmd-lang.html#command-lookup-order
|
| 1617 |
|
| 1618 | If the name isn't found, then `invoke` returns status 127.
|
| 1619 |
|
| 1620 | ---
|
| 1621 |
|
| 1622 | Run `invoke --show NAME` to see all categories a name is found in.
|
| 1623 |
|
| 1624 | - The `--show` flag respects the [command-lookup-order][]
|
| 1625 | - Shell keywords and aliases are shown, but `invoke` doesn't run them.
|
| 1626 |
|
| 1627 | ---
|
| 1628 |
|
| 1629 | Examples:
|
| 1630 |
|
| 1631 | invoke ls # usage error: no namespace flags
|
| 1632 |
|
| 1633 | invoke --builtin echo hi # like builtin echo hi
|
| 1634 | invoke --builtin --extern ls /tmp # like command ls /tmp (no function lookup)
|
| 1635 |
|
| 1636 | invoke --show true sleep ls # similar to type -a true sleep ls
|
| 1637 |
|
| 1638 | Related:
|
| 1639 |
|
| 1640 | - [builtin][] - like `--builtin`
|
| 1641 | - [command][] - like `--builtin --extern`
|
| 1642 | - [runproc][] - like `--proc --sh-func`
|
| 1643 | - [type][cmd/type] - like `--show`
|
| 1644 |
|
| 1645 | [builtin]: chap-builtin-cmd.html#builtin
|
| 1646 | [command]: chap-builtin-cmd.html#command
|
| 1647 | [runproc]: chap-builtin-cmd.html#runproc
|
| 1648 | [cmd/type]: chap-builtin-cmd.html#cmd/type
|
| 1649 | [command-lookup-order]: chap-cmd-lang.html#command-lookup-order
|
| 1650 |
|
| 1651 | ### runproc
|
| 1652 |
|
| 1653 | Runs a named proc with the given arguments. It's often useful as the only top
|
| 1654 | level statement in a "task file":
|
| 1655 |
|
| 1656 | proc p {
|
| 1657 | echo hi
|
| 1658 | }
|
| 1659 | runproc @ARGV
|
| 1660 |
|
| 1661 | Like 'builtin' and 'command', it affects the lookup of the first word.
|
| 1662 |
|
| 1663 | ### command
|
| 1664 |
|
| 1665 | command FLAG* CMD ARG*
|
| 1666 |
|
| 1667 | Look up CMD as a shell builtin or executable file, and execute it with the
|
| 1668 | given ARGs.
|
| 1669 |
|
| 1670 | Flags:
|
| 1671 |
|
| 1672 | -v Instead of executing CMD, print a description of it.
|
| 1673 | <!-- -p Use a default value for PATH that is guaranteed to find all of the
|
| 1674 | standard utilities.
|
| 1675 | -V Print a more verbose description of CMD.-->
|
| 1676 |
|
| 1677 | Note: [`invoke --show`][invoke] is more general than `command -v`.
|
| 1678 |
|
| 1679 | [invoke]: chap-builtin-cmd.html#invoke
|
| 1680 |
|
| 1681 | ### builtin
|
| 1682 |
|
| 1683 | builtin CMD ARG*
|
| 1684 |
|
| 1685 | Look up CMD as a shell builtin, and execute it with the given ARGs.
|
| 1686 |
|
| 1687 | ## Interactive
|
| 1688 |
|
| 1689 | ### alias
|
| 1690 |
|
| 1691 | alias NAME=CODE
|
| 1692 |
|
| 1693 | Make NAME a shortcut for executing CODE, e.g. `alias hi='echo hello'`.
|
| 1694 |
|
| 1695 | alias NAME
|
| 1696 |
|
| 1697 | Show the value of this alias.
|
| 1698 |
|
| 1699 | alias
|
| 1700 |
|
| 1701 | Show a list of all aliases.
|
| 1702 |
|
| 1703 | Tips:
|
| 1704 |
|
| 1705 | Prefer shell functions like:
|
| 1706 |
|
| 1707 | ls() {
|
| 1708 | command ls --color "$@"
|
| 1709 | }
|
| 1710 |
|
| 1711 | to aliases like:
|
| 1712 |
|
| 1713 | alias ls='ls --color'
|
| 1714 |
|
| 1715 | Functions are less likely to cause parsing problems.
|
| 1716 |
|
| 1717 | - Quoting like `\ls` or `'ls'` disables alias expansion
|
| 1718 | - To remove an existing alias, use [unalias](chap-builtin-cmd.html#unalias).
|
| 1719 |
|
| 1720 | ### unalias
|
| 1721 |
|
| 1722 | unalias NAME
|
| 1723 |
|
| 1724 | Remove the alias NAME.
|
| 1725 |
|
| 1726 | <!--Flag:
|
| 1727 |
|
| 1728 | -a Removes all existing aliases.-->
|
| 1729 |
|
| 1730 | ### history
|
| 1731 |
|
| 1732 | history FLAG*
|
| 1733 |
|
| 1734 | Display and manipulate the shell's history entries.
|
| 1735 |
|
| 1736 | history NUM
|
| 1737 |
|
| 1738 | Show the last NUM history entries.
|
| 1739 |
|
| 1740 | Flags:
|
| 1741 |
|
| 1742 | -c Clears the history.
|
| 1743 | -d POS Deletes the history entry at position POS.
|
| 1744 | <!-- -a
|
| 1745 | -n
|
| 1746 | -r
|
| 1747 | -w
|
| 1748 | -p
|
| 1749 | -s -->
|
| 1750 |
|
| 1751 | ### fc
|
| 1752 |
|
| 1753 | fc FLAG* FIRST? LAST?
|
| 1754 |
|
| 1755 | "Fix a command" from the shell's history.
|
| 1756 |
|
| 1757 | `fc -l` displays commands. FIRST and LAST specify a range of command numbers,
|
| 1758 | where:
|
| 1759 |
|
| 1760 | - A positive number is an index into the history list.
|
| 1761 | - A negative number is an offset from the current command.
|
| 1762 | - If FIRST is omitted, the value `-16` is used.
|
| 1763 | - If LAST is omitted, the current command is used.
|
| 1764 |
|
| 1765 | Flags:
|
| 1766 |
|
| 1767 | -l List commands (rather than editing)
|
| 1768 | -n Omit line numbers
|
| 1769 | -r Use reverse order (newest first)
|
| 1770 |
|
| 1771 | <!--
|
| 1772 | Not implemented
|
| 1773 |
|
| 1774 | -e EDITOR
|
| 1775 | -s
|
| 1776 | -->
|
| 1777 |
|
| 1778 | ## Unsupported
|
| 1779 |
|
| 1780 | ### enable
|
| 1781 |
|
| 1782 | Bash has this, but OSH won't implement it.
|
| 1783 |
|