| 1 |
 |
| 2 |
|
| 3 |
# TOML |
| 4 |
|
| 5 |
Tom's Obvious, Minimal Language. |
| 6 |
|
| 7 |
By Tom Preston-Werner, Pradyun Gedam, et al. |
| 8 |
|
| 9 |
## Objectives |
| 10 |
|
| 11 |
TOML aims to be a minimal configuration file format that's easy to read due to |
| 12 |
obvious semantics. TOML is designed to map unambiguously to a hash table. TOML |
| 13 |
should be easy to parse into data structures in a wide variety of languages. |
| 14 |
|
| 15 |
## Table of contents |
| 16 |
|
| 17 |
- [Preliminaries](#user-content-preliminaries) |
| 18 |
- [Comment](#user-content-comment) |
| 19 |
- [Key/Value Pair](#user-content-keyvalue-pair) |
| 20 |
- [Keys](#user-content-keys) |
| 21 |
- [String](#user-content-string) |
| 22 |
- [Integer](#user-content-integer) |
| 23 |
- [Float](#user-content-float) |
| 24 |
- [Boolean](#user-content-boolean) |
| 25 |
- [Offset Date-Time](#user-content-offset-date-time) |
| 26 |
- [Local Date-Time](#user-content-local-date-time) |
| 27 |
- [Local Date](#user-content-local-date) |
| 28 |
- [Local Time](#user-content-local-time) |
| 29 |
- [Array](#user-content-array) |
| 30 |
- [Table](#user-content-table) |
| 31 |
- [Inline Table](#user-content-inline-table) |
| 32 |
- [Array of Tables](#user-content-array-of-tables) |
| 33 |
- [Filename Extension](#user-content-filename-extension) |
| 34 |
- [MIME Type](#user-content-mime-type) |
| 35 |
- [ABNF Grammar](#user-content-abnf-grammar) |
| 36 |
|
| 37 |
## Preliminaries |
| 38 |
|
| 39 |
- TOML is case-sensitive. |
| 40 |
- Whitespace means tab (U+0009) or space (U+0020). |
| 41 |
- Newline means LF (U+000A) or CRLF (U+000D U+000A). |
| 42 |
- A TOML file must be a valid UTF-8 encoded Unicode document. |
| 43 |
|
| 44 |
Specifically this means that a file _as a whole_ must form a |
| 45 |
[well-formed code-unit sequence](https://unicode.org/glossary/#well_formed_code_unit_sequence). |
| 46 |
Otherwise, it must be rejected (preferably) or have ill-formed byte sequences |
| 47 |
replaced with U+FFFD, as per the Unicode specification. |
| 48 |
|
| 49 |
## Comment |
| 50 |
|
| 51 |
A hash symbol marks the rest of the line as a comment, except when inside a |
| 52 |
string. |
| 53 |
|
| 54 |
```toml |
| 55 |
# This is a full-line comment |
| 56 |
key = "value" # This is a comment at the end of a line |
| 57 |
another = "# This is not a comment" |
| 58 |
``` |
| 59 |
|
| 60 |
Control characters other than tab (U+0000 to U+0008, U+000A to U+001F, U+007F) |
| 61 |
are not permitted in comments. |
| 62 |
|
| 63 |
Comments should be used to communicate between the human readers of a file. |
| 64 |
Parsers must not modify keys or values, based on the presence (or contents) of a |
| 65 |
comment. |
| 66 |
|
| 67 |
## Key/Value Pair |
| 68 |
|
| 69 |
The primary building block of a TOML document is the key/value pair. |
| 70 |
|
| 71 |
Keys are on the left of the equals sign and values are on the right. Whitespace |
| 72 |
is ignored around key names and values. The key, equals sign, and value must be |
| 73 |
on the same line (though some values can be broken over multiple lines). |
| 74 |
|
| 75 |
```toml |
| 76 |
key = "value" |
| 77 |
``` |
| 78 |
|
| 79 |
Values must have one of the following types. |
| 80 |
|
| 81 |
- [String](#user-content-string) |
| 82 |
- [Integer](#user-content-integer) |
| 83 |
- [Float](#user-content-float) |
| 84 |
- [Boolean](#user-content-boolean) |
| 85 |
- [Offset Date-Time](#user-content-offset-date-time) |
| 86 |
- [Local Date-Time](#user-content-local-date-time) |
| 87 |
- [Local Date](#user-content-local-date) |
| 88 |
- [Local Time](#user-content-local-time) |
| 89 |
- [Array](#user-content-array) |
| 90 |
- [Inline Table](#user-content-inline-table) |
| 91 |
|
| 92 |
Unspecified values are invalid. |
| 93 |
|
| 94 |
```toml |
| 95 |
key = # INVALID |
| 96 |
``` |
| 97 |
|
| 98 |
There must be a newline (or EOF) after a key/value pair. (See |
| 99 |
[Inline Table](#user-content-inline-table) for exceptions.) |
| 100 |
|
| 101 |
``` |
| 102 |
first = "Tom" last = "Preston-Werner" # INVALID |
| 103 |
``` |
| 104 |
|
| 105 |
## Keys |
| 106 |
|
| 107 |
A key may be either bare, quoted, or dotted. |
| 108 |
|
| 109 |
**Bare keys** may only contain ASCII letters, ASCII digits, underscores, and |
| 110 |
dashes (`A-Za-z0-9_-`). Note that bare keys are allowed to be composed of only |
| 111 |
ASCII digits, e.g. `1234`, but are always interpreted as strings. |
| 112 |
|
| 113 |
ℹ️ The exact ranges of allowed code points can be found in the |
| 114 |
[ABNF grammar file][abnf]. |
| 115 |
|
| 116 |
```toml |
| 117 |
key = "value" |
| 118 |
bare_key = "value" |
| 119 |
bare-key = "value" |
| 120 |
1234 = "value" |
| 121 |
``` |
| 122 |
|
| 123 |
**Quoted keys** follow the exact same rules as either basic strings or literal |
| 124 |
strings and allow you to use a much broader set of key names. Best practice is |
| 125 |
to use bare keys except when absolutely necessary. |
| 126 |
|
| 127 |
```toml |
| 128 |
"127.0.0.1" = "value" |
| 129 |
"character encoding" = "value" |
| 130 |
"ʎǝʞ" = "value" |
| 131 |
'key2' = "value" |
| 132 |
'quoted "value"' = "value" |
| 133 |
``` |
| 134 |
|
| 135 |
A bare key must be non-empty, but an empty quoted key is allowed (though |
| 136 |
discouraged). You cannot use multi-line strings to define quoted keys. |
| 137 |
|
| 138 |
```toml |
| 139 |
= "no key name" # INVALID |
| 140 |
"""key""" = "not allowed" # INVALID |
| 141 |
"" = "blank" # VALID but discouraged |
| 142 |
'' = 'blank' # VALID but discouraged |
| 143 |
``` |
| 144 |
|
| 145 |
**Dotted keys** are a sequence of bare or quoted keys joined with a dot. This |
| 146 |
allows for grouping similar properties together: |
| 147 |
|
| 148 |
```toml |
| 149 |
name = "Orange" |
| 150 |
physical.color = "orange" |
| 151 |
physical.shape = "round" |
| 152 |
site."google.com" = true |
| 153 |
``` |
| 154 |
|
| 155 |
In JSON land, that would give you the following structure: |
| 156 |
|
| 157 |
```json |
| 158 |
{ |
| 159 |
"name": "Orange", |
| 160 |
"physical": { |
| 161 |
"color": "orange", |
| 162 |
"shape": "round" |
| 163 |
}, |
| 164 |
"site": { |
| 165 |
"google.com": true |
| 166 |
} |
| 167 |
} |
| 168 |
``` |
| 169 |
|
| 170 |
For details regarding the tables that dotted keys define, refer to the |
| 171 |
[Table](#user-content-table) section below. |
| 172 |
|
| 173 |
Whitespace around dot-separated parts is ignored. However, best practice is to |
| 174 |
not use any extraneous whitespace. |
| 175 |
|
| 176 |
```toml |
| 177 |
fruit.name = "banana" # this is best practice |
| 178 |
fruit. color = "yellow" # same as fruit.color |
| 179 |
fruit . flavor = "banana" # same as fruit.flavor |
| 180 |
``` |
| 181 |
|
| 182 |
Indentation is treated as whitespace and ignored. |
| 183 |
|
| 184 |
Defining a key multiple times is invalid. |
| 185 |
|
| 186 |
``` |
| 187 |
# DO NOT DO THIS |
| 188 |
name = "Tom" |
| 189 |
name = "Pradyun" |
| 190 |
``` |
| 191 |
|
| 192 |
Note that bare keys and quoted keys are equivalent: |
| 193 |
|
| 194 |
``` |
| 195 |
# THIS WILL NOT WORK |
| 196 |
spelling = "favorite" |
| 197 |
"spelling" = "favourite" |
| 198 |
``` |
| 199 |
|
| 200 |
As long as a key hasn't been directly defined, you may still write to it and to |
| 201 |
names within it. |
| 202 |
|
| 203 |
```toml |
| 204 |
# This makes the key "fruit" into a table. |
| 205 |
fruit.apple.smooth = true |
| 206 |
|
| 207 |
# So then you can add to the table "fruit" like so: |
| 208 |
fruit.orange = 2 |
| 209 |
``` |
| 210 |
|
| 211 |
``` |
| 212 |
# THE FOLLOWING IS INVALID |
| 213 |
|
| 214 |
# This defines the value of fruit.apple to be an integer. |
| 215 |
fruit.apple = 1 |
| 216 |
|
| 217 |
# But then this treats fruit.apple like it's a table. |
| 218 |
# You can't turn an integer into a table. |
| 219 |
fruit.apple.smooth = true |
| 220 |
``` |
| 221 |
|
| 222 |
Defining dotted keys out-of-order is discouraged. |
| 223 |
|
| 224 |
```toml |
| 225 |
# VALID BUT DISCOURAGED |
| 226 |
|
| 227 |
apple.type = "fruit" |
| 228 |
orange.type = "fruit" |
| 229 |
|
| 230 |
apple.skin = "thin" |
| 231 |
orange.skin = "thick" |
| 232 |
|
| 233 |
apple.color = "red" |
| 234 |
orange.color = "orange" |
| 235 |
``` |
| 236 |
|
| 237 |
```toml |
| 238 |
# RECOMMENDED |
| 239 |
|
| 240 |
apple.type = "fruit" |
| 241 |
apple.skin = "thin" |
| 242 |
apple.color = "red" |
| 243 |
|
| 244 |
orange.type = "fruit" |
| 245 |
orange.skin = "thick" |
| 246 |
orange.color = "orange" |
| 247 |
``` |
| 248 |
|
| 249 |
Since bare keys can be composed of only ASCII integers, it is possible to write |
| 250 |
dotted keys that look like floats but are 2-part dotted keys. Don't do this |
| 251 |
unless you have a good reason to (you probably don't). |
| 252 |
|
| 253 |
```toml |
| 254 |
3.14159 = "pi" |
| 255 |
``` |
| 256 |
|
| 257 |
The above TOML maps to the following JSON. |
| 258 |
|
| 259 |
```json |
| 260 |
{ "3": { "14159": "pi" } } |
| 261 |
``` |
| 262 |
|
| 263 |
## String |
| 264 |
|
| 265 |
There are four ways to express strings: basic, multi-line basic, literal, and |
| 266 |
multi-line literal. All strings must contain only Unicode characters. |
| 267 |
|
| 268 |
**Basic strings** are surrounded by quotation marks (`"`). Any Unicode character |
| 269 |
may be used except those that must be escaped: quotation mark, backslash, and |
| 270 |
the control characters other than tab (U+0000 to U+0008, U+000A to U+001F, |
| 271 |
U+007F). |
| 272 |
|
| 273 |
```toml |
| 274 |
str = "I'm a string. \"You can quote me\". Name\tJos\xE9\nLocation\tSF." |
| 275 |
``` |
| 276 |
|
| 277 |
For convenience, some popular characters have a compact escape sequence. |
| 278 |
|
| 279 |
``` |
| 280 |
\b - backspace (U+0008) |
| 281 |
\t - tab (U+0009) |
| 282 |
\n - linefeed (U+000A) |
| 283 |
\f - form feed (U+000C) |
| 284 |
\r - carriage return (U+000D) |
| 285 |
\e - escape (U+001B) |
| 286 |
\" - quote (U+0022) |
| 287 |
\\ - backslash (U+005C) |
| 288 |
\xHH - unicode (U+00HH) |
| 289 |
\uHHHH - unicode (U+HHHH) |
| 290 |
\UHHHHHHHH - unicode (U+HHHHHHHH) |
| 291 |
``` |
| 292 |
|
| 293 |
Any Unicode character may be escaped with the `\xHH`, `\uHHHH`, or `\UHHHHHHHH` |
| 294 |
forms. The escape codes must be Unicode |
| 295 |
[scalar values](https://unicode.org/glossary/#unicode_scalar_value). |
| 296 |
|
| 297 |
Keep in mind that all TOML strings are sequences of Unicode characters, _not_ |
| 298 |
byte sequences. For binary data, avoid using these escape codes. Instead, |
| 299 |
external binary-to-text encoding strategies, like hexadecimal sequences or |
| 300 |
[Base64](https://www.base64decode.org/), are recommended for converting between |
| 301 |
bytes and strings. |
| 302 |
|
| 303 |
All other escape sequences not listed above are reserved; if they are used, TOML |
| 304 |
should produce an error. |
| 305 |
|
| 306 |
Sometimes you need to express passages of text (e.g. translation files) or would |
| 307 |
like to break up a very long string into multiple lines. TOML makes this easy. |
| 308 |
|
| 309 |
**Multi-line basic strings** are surrounded by three quotation marks on each |
| 310 |
side and allow newlines. A newline immediately following the opening delimiter |
| 311 |
will be trimmed. All other whitespace and newline characters remain intact. |
| 312 |
|
| 313 |
```toml |
| 314 |
str1 = """ |
| 315 |
Roses are red |
| 316 |
Violets are blue""" |
| 317 |
``` |
| 318 |
|
| 319 |
TOML parsers should feel free to normalize newline to whatever makes sense for |
| 320 |
their platform. |
| 321 |
|
| 322 |
```toml |
| 323 |
# On a Unix system, the above multi-line string will most likely be the same as: |
| 324 |
str2 = "Roses are red\nViolets are blue" |
| 325 |
|
| 326 |
# On a Windows system, it will most likely be equivalent to: |
| 327 |
str3 = "Roses are red\r\nViolets are blue" |
| 328 |
``` |
| 329 |
|
| 330 |
For writing long strings without introducing extraneous whitespace, use a "line |
| 331 |
ending backslash". When the last non-whitespace character on a line is an |
| 332 |
unescaped `\`, it will be trimmed along with all whitespace (including newlines) |
| 333 |
up to the next non-whitespace character or closing delimiter. All of the escape |
| 334 |
sequences that are valid for basic strings are also valid for multi-line basic |
| 335 |
strings. |
| 336 |
|
| 337 |
```toml |
| 338 |
# The following strings are byte-for-byte equivalent: |
| 339 |
str1 = "The quick brown fox jumps over the lazy dog." |
| 340 |
|
| 341 |
str2 = """ |
| 342 |
The quick brown \ |
| 343 |
|
| 344 |
|
| 345 |
fox jumps over \ |
| 346 |
the lazy dog.""" |
| 347 |
|
| 348 |
str3 = """\ |
| 349 |
The quick brown \ |
| 350 |
fox jumps over \ |
| 351 |
the lazy dog.\ |
| 352 |
""" |
| 353 |
``` |
| 354 |
|
| 355 |
Any Unicode character may be used except those that must be escaped: backslash |
| 356 |
and the control characters other than tab, line feed, and carriage return |
| 357 |
(U+0000 to U+0008, U+000B, U+000C, U+000E to U+001F, U+007F). Carriage returns |
| 358 |
(U+000D) are only allowed as part of a newline sequence. |
| 359 |
|
| 360 |
You can write a quotation mark, or two adjacent quotation marks, anywhere inside |
| 361 |
a multi-line basic string. They can also be written just inside the delimiters. |
| 362 |
|
| 363 |
```toml |
| 364 |
str4 = """Here are two quotation marks: "". Simple enough.""" |
| 365 |
# str5 = """Here are three quotation marks: """.""" # INVALID |
| 366 |
str5 = """Here are three quotation marks: ""\".""" |
| 367 |
str6 = """Here are fifteen quotation marks: ""\"""\"""\"""\"""\".""" |
| 368 |
|
| 369 |
# "This," she said, "is just a pointless statement." |
| 370 |
str7 = """"This," she said, "is just a pointless statement."""" |
| 371 |
``` |
| 372 |
|
| 373 |
If you're a frequent specifier of Windows paths or regular expressions, then |
| 374 |
having to escape backslashes quickly becomes tedious and error-prone. To help, |
| 375 |
TOML supports literal strings which do not allow escaping at all. |
| 376 |
|
| 377 |
**Literal strings** are surrounded by single quotes. Like basic strings, they |
| 378 |
must appear on a single line: |
| 379 |
|
| 380 |
```toml |
| 381 |
# What you see is what you get. |
| 382 |
winpath = 'C:\Users\nodejs\templates' |
| 383 |
winpath2 = '\\ServerX\admin$\system32\' |
| 384 |
quoted = 'Tom "Dubs" Preston-Werner' |
| 385 |
regex = '<\i\c*\s*>' |
| 386 |
``` |
| 387 |
|
| 388 |
Since there is no escaping, there is no way to write a single quote inside a |
| 389 |
literal string enclosed by single quotes. Luckily, TOML supports a multi-line |
| 390 |
version of literal strings that solves this problem. |
| 391 |
|
| 392 |
**Multi-line literal strings** are surrounded by three single quotes on each |
| 393 |
side and allow newlines. Like literal strings, there is no escaping whatsoever. |
| 394 |
A newline immediately following the opening delimiter will be trimmed. TOML |
| 395 |
parsers must normalize newlines in the same manner as multi-line basic strings. |
| 396 |
|
| 397 |
All other content between the delimiters is interpreted as-is without |
| 398 |
modification. |
| 399 |
|
| 400 |
```toml |
| 401 |
regex2 = '''I [dw]on't need \d{2} apples''' |
| 402 |
lines = ''' |
| 403 |
The first newline is |
| 404 |
trimmed in literal strings. |
| 405 |
All other whitespace |
| 406 |
is preserved. |
| 407 |
''' |
| 408 |
``` |
| 409 |
|
| 410 |
You can write 1 or 2 single quotes anywhere within a multi-line literal string, |
| 411 |
but sequences of three or more single quotes are not permitted. |
| 412 |
|
| 413 |
```toml |
| 414 |
quot15 = '''Here are fifteen quotation marks: """""""""""""""''' |
| 415 |
|
| 416 |
# apos15 = '''Here are fifteen apostrophes: '''''''''''''''''' # INVALID |
| 417 |
apos15 = "Here are fifteen apostrophes: '''''''''''''''" |
| 418 |
|
| 419 |
# 'That,' she said, 'is still pointless.' |
| 420 |
str = ''''That,' she said, 'is still pointless.'''' |
| 421 |
``` |
| 422 |
|
| 423 |
Control characters other than tab are not permitted in a literal string. |
| 424 |
|
| 425 |
## Integer |
| 426 |
|
| 427 |
Integers are whole numbers. Positive numbers may be prefixed with a plus sign. |
| 428 |
Negative numbers are prefixed with a minus sign. |
| 429 |
|
| 430 |
```toml |
| 431 |
int1 = +99 |
| 432 |
int2 = 42 |
| 433 |
int3 = 0 |
| 434 |
int4 = -17 |
| 435 |
``` |
| 436 |
|
| 437 |
For large numbers, you may use underscores between digits to enhance |
| 438 |
readability. Each underscore must be surrounded by at least one digit on each |
| 439 |
side. |
| 440 |
|
| 441 |
```toml |
| 442 |
int5 = 1_000 |
| 443 |
int6 = 5_349_221 |
| 444 |
int7 = 53_49_221 # Indian number system grouping |
| 445 |
int8 = 1_2_3_4_5 # VALID but discouraged |
| 446 |
``` |
| 447 |
|
| 448 |
Leading zeros are not allowed. Integer values `-0` and `+0` are valid and |
| 449 |
identical to an unprefixed zero. |
| 450 |
|
| 451 |
Non-negative integer values may also be expressed in hexadecimal, octal, or |
| 452 |
binary. In these formats, leading `+` is not allowed and leading zeros are |
| 453 |
allowed (after the prefix). Hex values are case-insensitive. Underscores are |
| 454 |
allowed between digits (but not between the prefix and the value). |
| 455 |
|
| 456 |
```toml |
| 457 |
# hexadecimal with prefix `0x` |
| 458 |
hex1 = 0xDEADBEEF |
| 459 |
hex2 = 0xdeadbeef |
| 460 |
hex3 = 0xdead_beef |
| 461 |
|
| 462 |
# octal with prefix `0o` |
| 463 |
oct1 = 0o01234567 |
| 464 |
oct2 = 0o755 # useful for Unix file permissions |
| 465 |
|
| 466 |
# binary with prefix `0b` |
| 467 |
bin1 = 0b11010110 |
| 468 |
``` |
| 469 |
|
| 470 |
Arbitrary 64-bit signed integers (from −2^63 to 2^63−1) should be accepted and |
| 471 |
handled losslessly. If an integer cannot be represented losslessly, an error |
| 472 |
must be thrown. |
| 473 |
|
| 474 |
## Float |
| 475 |
|
| 476 |
Floats should be implemented as IEEE 754 binary64 values. |
| 477 |
|
| 478 |
A float consists of an integer part (which follows the same rules as decimal |
| 479 |
integer values) followed by a fractional part and/or an exponent part. If both a |
| 480 |
fractional part and exponent part are present, the fractional part must precede |
| 481 |
the exponent part. |
| 482 |
|
| 483 |
```toml |
| 484 |
# fractional |
| 485 |
flt1 = +1.0 |
| 486 |
flt2 = 3.1415 |
| 487 |
flt3 = -0.01 |
| 488 |
|
| 489 |
# exponent |
| 490 |
flt4 = 5e+22 |
| 491 |
flt5 = 1e06 |
| 492 |
flt6 = -2E-2 |
| 493 |
|
| 494 |
# both |
| 495 |
flt7 = 6.626e-34 |
| 496 |
``` |
| 497 |
|
| 498 |
A fractional part is a decimal point followed by one or more digits. |
| 499 |
|
| 500 |
An exponent part is an E (upper or lower case) followed by an integer part |
| 501 |
(which follows the same rules as decimal integer values but may include leading |
| 502 |
zeros). |
| 503 |
|
| 504 |
The decimal point, if used, must be surrounded by at least one digit on each |
| 505 |
side. |
| 506 |
|
| 507 |
``` |
| 508 |
# INVALID FLOATS |
| 509 |
invalid_float_1 = .7 |
| 510 |
invalid_float_2 = 7. |
| 511 |
invalid_float_3 = 3.e+20 |
| 512 |
``` |
| 513 |
|
| 514 |
Similar to integers, you may use underscores to enhance readability. Each |
| 515 |
underscore must be surrounded by at least one digit. |
| 516 |
|
| 517 |
```toml |
| 518 |
flt8 = 224_617.445_991_228 |
| 519 |
``` |
| 520 |
|
| 521 |
Float values `-0.0` and `+0.0` are valid and should map according to IEEE 754. |
| 522 |
|
| 523 |
Special float values can also be expressed. They are always lowercase. |
| 524 |
|
| 525 |
```toml |
| 526 |
# infinity |
| 527 |
sf1 = inf # positive infinity |
| 528 |
sf2 = +inf # positive infinity |
| 529 |
sf3 = -inf # negative infinity |
| 530 |
|
| 531 |
# not a number |
| 532 |
sf4 = nan # actual sNaN/qNaN encoding is implementation-specific |
| 533 |
sf5 = +nan # same as `nan` |
| 534 |
sf6 = -nan # valid, actual encoding is implementation-specific |
| 535 |
``` |
| 536 |
|
| 537 |
## Boolean |
| 538 |
|
| 539 |
Booleans are just the tokens you're used to. Always lowercase. |
| 540 |
|
| 541 |
```toml |
| 542 |
bool1 = true |
| 543 |
bool2 = false |
| 544 |
``` |
| 545 |
|
| 546 |
## Offset Date-Time |
| 547 |
|
| 548 |
To unambiguously represent a specific instant in time, you may use an |
| 549 |
[RFC 3339](https://tools.ietf.org/html/rfc3339) formatted date-time with offset. |
| 550 |
|
| 551 |
```toml |
| 552 |
odt1 = 1979-05-27T07:32:00Z |
| 553 |
odt2 = 1979-05-27T00:32:00-07:00 |
| 554 |
odt3 = 1979-05-27T00:32:00.5-07:00 |
| 555 |
odt4 = 1979-05-27T00:32:00.999999-07:00 |
| 556 |
``` |
| 557 |
|
| 558 |
For the sake of readability, you may replace the T delimiter between date and |
| 559 |
time with a space character (as permitted by RFC 3339 section 5.6). |
| 560 |
|
| 561 |
```toml |
| 562 |
odt4 = 1979-05-27 07:32:00Z |
| 563 |
``` |
| 564 |
|
| 565 |
One exception to RFC 3339 is permitted: seconds may be omitted, in which case |
| 566 |
`:00` will be assumed. The offset immediately follows the minutes. |
| 567 |
|
| 568 |
```toml |
| 569 |
odt5 = 1979-05-27 07:32Z |
| 570 |
odt6 = 1979-05-27 07:32-07:00 |
| 571 |
``` |
| 572 |
|
| 573 |
Implementations are required to support at least millisecond precision. |
| 574 |
Additional digits of precision may be specified, but if they exceed the |
| 575 |
supported precision then the extra digits must be truncated, not rounded. |
| 576 |
|
| 577 |
## Local Date-Time |
| 578 |
|
| 579 |
If you omit the offset from an [RFC 3339](https://tools.ietf.org/html/rfc3339) |
| 580 |
formatted date-time, it will represent the given date-time without any relation |
| 581 |
to an offset or timezone. It cannot be converted to an instant in time without |
| 582 |
additional information. Conversion to an instant, if required, is |
| 583 |
implementation-specific. |
| 584 |
|
| 585 |
```toml |
| 586 |
ldt1 = 1979-05-27T07:32:00 |
| 587 |
ldt2 = 1979-05-27T07:32:00.5 |
| 588 |
ldt3 = 1979-05-27T00:32:00.999999 |
| 589 |
``` |
| 590 |
|
| 591 |
Seconds may be omitted, in which case `:00` will be assumed. |
| 592 |
|
| 593 |
```toml |
| 594 |
ldt3 = 1979-05-27T07:32 |
| 595 |
``` |
| 596 |
|
| 597 |
Millisecond precision is required. Further precision of fractional seconds is |
| 598 |
implementation-specific. If the value contains greater precision than the |
| 599 |
implementation can support, the additional precision must be truncated, not |
| 600 |
rounded. |
| 601 |
|
| 602 |
Implementations are required to support at least millisecond precision. |
| 603 |
Additional digits of precision may be specified, but if they exceed the |
| 604 |
supported precision then the extra digits must be truncated, not rounded. |
| 605 |
|
| 606 |
## Local Date |
| 607 |
|
| 608 |
If you include only the date portion of an |
| 609 |
[RFC 3339](https://tools.ietf.org/html/rfc3339) formatted date-time, it will |
| 610 |
represent that entire day without any relation to an offset or timezone. |
| 611 |
|
| 612 |
```toml |
| 613 |
ld1 = 1979-05-27 |
| 614 |
``` |
| 615 |
|
| 616 |
## Local Time |
| 617 |
|
| 618 |
If you include only the time portion of an |
| 619 |
[RFC 3339](https://tools.ietf.org/html/rfc3339) formatted date-time, it will |
| 620 |
represent that time of day without any relation to a specific day or any offset |
| 621 |
or timezone. |
| 622 |
|
| 623 |
```toml |
| 624 |
lt1 = 07:32:00 |
| 625 |
lt2 = 00:32:00.5 |
| 626 |
lt3 = 00:32:00.999999 |
| 627 |
``` |
| 628 |
|
| 629 |
Seconds may be omitted, in which case `:00` will be assumed. |
| 630 |
|
| 631 |
```toml |
| 632 |
lt3 = 07:32 |
| 633 |
``` |
| 634 |
|
| 635 |
Millisecond precision is required. Further precision of fractional seconds is |
| 636 |
implementation-specific. If the value contains greater precision than the |
| 637 |
implementation can support, the additional precision must be truncated, not |
| 638 |
rounded. |
| 639 |
|
| 640 |
Implementations are required to support at least millisecond precision. |
| 641 |
Additional digits of precision may be specified, but if they exceed the |
| 642 |
supported precision then the extra digits must be truncated, not rounded. |
| 643 |
|
| 644 |
## Array |
| 645 |
|
| 646 |
Arrays are ordered values surrounded by square brackets. Whitespace is ignored. |
| 647 |
Elements are separated by commas. Arrays can contain values of the same data |
| 648 |
types as allowed in key/value pairs. Values of different types may be mixed. |
| 649 |
|
| 650 |
```toml |
| 651 |
integers = [ 1, 2, 3 ] |
| 652 |
colors = [ "red", "yellow", "green" ] |
| 653 |
nested_arrays_of_ints = [ [ 1, 2 ], [3, 4, 5] ] |
| 654 |
nested_mixed_array = [ [ 1, 2 ], ["a", "b", "c"] ] |
| 655 |
string_array = [ "all", 'strings', """are the same""", '''type''' ] |
| 656 |
|
| 657 |
# Mixed-type arrays are allowed |
| 658 |
numbers = [ 0.1, 0.2, 0.5, 1, 2, 5 ] |
| 659 |
contributors = [ |
| 660 |
"Foo Bar <[email protected]>", |
| 661 |
{ name = "Baz Qux", email = "[email protected]", url = "https://example.com/bazqux" } |
| 662 |
] |
| 663 |
``` |
| 664 |
|
| 665 |
Arrays can span multiple lines. A terminating comma (also called a trailing |
| 666 |
comma) is permitted after the last value of the array. Any number of newlines |
| 667 |
and comments may precede values, commas, and the closing bracket. Indentation |
| 668 |
between array values and commas is treated as whitespace and ignored. |
| 669 |
|
| 670 |
```toml |
| 671 |
integers2 = [ |
| 672 |
1, 2, 3 |
| 673 |
] |
| 674 |
|
| 675 |
integers3 = [ |
| 676 |
1, |
| 677 |
2, # this is ok |
| 678 |
] |
| 679 |
``` |
| 680 |
|
| 681 |
## Table |
| 682 |
|
| 683 |
Tables (also known as hash tables or dictionaries) are collections of key/value |
| 684 |
pairs. They are defined by headers, with square brackets on a line by |
| 685 |
themselves. You can tell headers apart from arrays because arrays are only ever |
| 686 |
values. |
| 687 |
|
| 688 |
```toml |
| 689 |
[table] |
| 690 |
``` |
| 691 |
|
| 692 |
Under that, and until the next header or EOF, are the key/values of that table. |
| 693 |
Key/value pairs within tables are not guaranteed to be in any specific order. |
| 694 |
|
| 695 |
```toml |
| 696 |
[table-1] |
| 697 |
key1 = "some string" |
| 698 |
key2 = 123 |
| 699 |
|
| 700 |
[table-2] |
| 701 |
key1 = "another string" |
| 702 |
key2 = 456 |
| 703 |
``` |
| 704 |
|
| 705 |
Naming rules for tables are the same as for keys (see definition of |
| 706 |
[Keys](#user-content-keys) above). |
| 707 |
|
| 708 |
```toml |
| 709 |
[dog."tater.man"] |
| 710 |
type.name = "pug" |
| 711 |
``` |
| 712 |
|
| 713 |
In JSON land, that would give you the following structure: |
| 714 |
|
| 715 |
```json |
| 716 |
{ "dog": { "tater.man": { "type": { "name": "pug" } } } } |
| 717 |
``` |
| 718 |
|
| 719 |
Whitespace around the key is ignored. However, best practice is to not use any |
| 720 |
extraneous whitespace. |
| 721 |
|
| 722 |
```toml |
| 723 |
[a.b.c] # this is best practice |
| 724 |
[ d.e.f ] # same as [d.e.f] |
| 725 |
[ g . h . i ] # same as [g.h.i] |
| 726 |
[ j . "ʞ" . 'l' ] # same as [j."ʞ".'l'] |
| 727 |
``` |
| 728 |
|
| 729 |
Indentation is treated as whitespace and ignored. |
| 730 |
|
| 731 |
You don't need to specify all the super-tables if you don't want to. TOML knows |
| 732 |
how to do it for you. |
| 733 |
|
| 734 |
```toml |
| 735 |
# [x] you |
| 736 |
# [x.y] don't |
| 737 |
# [x.y.z] need these |
| 738 |
[x.y.z.w] # for this to work |
| 739 |
|
| 740 |
[x] # defining a super-table afterward is ok |
| 741 |
``` |
| 742 |
|
| 743 |
Empty tables are allowed and simply have no key/value pairs within them. |
| 744 |
|
| 745 |
Like keys, you cannot define a table more than once. Doing so is invalid. |
| 746 |
|
| 747 |
``` |
| 748 |
# DO NOT DO THIS |
| 749 |
|
| 750 |
[fruit] |
| 751 |
apple = "red" |
| 752 |
|
| 753 |
[fruit] |
| 754 |
orange = "orange" |
| 755 |
``` |
| 756 |
|
| 757 |
``` |
| 758 |
# DO NOT DO THIS EITHER |
| 759 |
|
| 760 |
[fruit] |
| 761 |
apple = "red" |
| 762 |
|
| 763 |
[fruit.apple] |
| 764 |
texture = "smooth" |
| 765 |
``` |
| 766 |
|
| 767 |
Defining tables out-of-order is discouraged. |
| 768 |
|
| 769 |
```toml |
| 770 |
# VALID BUT DISCOURAGED |
| 771 |
[fruit.apple] |
| 772 |
[animal] |
| 773 |
[fruit.orange] |
| 774 |
``` |
| 775 |
|
| 776 |
```toml |
| 777 |
# RECOMMENDED |
| 778 |
[fruit.apple] |
| 779 |
[fruit.orange] |
| 780 |
[animal] |
| 781 |
``` |
| 782 |
|
| 783 |
The top-level table, also called the root table, starts at the beginning of the |
| 784 |
document and ends just before the first table header (or EOF). Unlike other |
| 785 |
tables, it is nameless and cannot be relocated. |
| 786 |
|
| 787 |
```toml |
| 788 |
# Top-level table begins. |
| 789 |
name = "Fido" |
| 790 |
breed = "pug" |
| 791 |
|
| 792 |
# Top-level table ends. |
| 793 |
[owner] |
| 794 |
name = "Regina Dogman" |
| 795 |
member_since = 1999-08-04 |
| 796 |
``` |
| 797 |
|
| 798 |
Dotted keys create and define a table for each key part before the last one. Any |
| 799 |
such table must have all its key/value pairs defined under the current `[table]` |
| 800 |
header, or in the root table if defined before all headers, or in one inline |
| 801 |
table. |
| 802 |
|
| 803 |
```toml |
| 804 |
fruit.apple.color = "red" |
| 805 |
# Defines a table named fruit |
| 806 |
# Defines a table named fruit.apple |
| 807 |
|
| 808 |
fruit.apple.taste.sweet = true |
| 809 |
# Defines a table named fruit.apple.taste |
| 810 |
# fruit and fruit.apple were already created |
| 811 |
``` |
| 812 |
|
| 813 |
Since tables cannot be defined more than once, redefining such tables using a |
| 814 |
`[table]` header is not allowed. Likewise, using dotted keys to redefine tables |
| 815 |
already defined in `[table]` form is not allowed. The `[table]` form can, |
| 816 |
however, be used to define sub-tables within tables defined via dotted keys. |
| 817 |
|
| 818 |
```toml |
| 819 |
[fruit] |
| 820 |
apple.color = "red" |
| 821 |
apple.taste.sweet = true |
| 822 |
|
| 823 |
# [fruit.apple] # INVALID |
| 824 |
# [fruit.apple.taste] # INVALID |
| 825 |
|
| 826 |
[fruit.apple.texture] # you can add sub-tables |
| 827 |
smooth = true |
| 828 |
``` |
| 829 |
|
| 830 |
## Inline Table |
| 831 |
|
| 832 |
Inline tables provide a more compact syntax for expressing tables. They are |
| 833 |
especially useful for grouped nested data that can otherwise quickly become |
| 834 |
verbose. |
| 835 |
|
| 836 |
Inline tables are fully defined within curly braces: `{` and `}`. Within the |
| 837 |
braces, zero or more comma-separated key/value pairs may appear. Key/value pairs |
| 838 |
take the same form as key/value pairs in standard tables. All value types are |
| 839 |
allowed, including inline tables. |
| 840 |
|
| 841 |
Inline tables can have multiple key/value pairs on the same line, or they can be |
| 842 |
put on different lines. A terminating comma (also called trailing comma) is |
| 843 |
permitted after the last key/value pair. |
| 844 |
|
| 845 |
```toml |
| 846 |
name = { first = "Tom", last = "Preston-Werner" } |
| 847 |
point = {x=1, y=2} |
| 848 |
animal = { type.name = "pug" } |
| 849 |
contact = { |
| 850 |
personal = { |
| 851 |
name = "Donald Duck", |
| 852 |
email = "[email protected]", |
| 853 |
}, |
| 854 |
work = { |
| 855 |
name = "Coin cleaner", |
| 856 |
email = "[email protected]", |
| 857 |
}, |
| 858 |
} |
| 859 |
``` |
| 860 |
|
| 861 |
The inline tables above are identical to the following standard table |
| 862 |
definitions: |
| 863 |
|
| 864 |
```toml |
| 865 |
[name] |
| 866 |
first = "Tom" |
| 867 |
last = "Preston-Werner" |
| 868 |
|
| 869 |
[point] |
| 870 |
x = 1 |
| 871 |
y = 2 |
| 872 |
|
| 873 |
[animal] |
| 874 |
type.name = "pug" |
| 875 |
|
| 876 |
[contact.personal] |
| 877 |
name = "Donald Duck" |
| 878 |
email = "[email protected]" |
| 879 |
|
| 880 |
[contact.work] |
| 881 |
name = "Coin cleaner" |
| 882 |
email = "[email protected]" |
| 883 |
``` |
| 884 |
|
| 885 |
Inline tables are fully self-contained and define all keys and sub-tables within |
| 886 |
them. Keys and sub-tables cannot be added outside the braces. |
| 887 |
|
| 888 |
```toml |
| 889 |
[product] |
| 890 |
type = { name = "Nail" } |
| 891 |
# type.edible = false # INVALID |
| 892 |
``` |
| 893 |
|
| 894 |
Similarly, inline tables cannot be used to add keys or sub-tables to an |
| 895 |
already-defined table. |
| 896 |
|
| 897 |
```toml |
| 898 |
[product] |
| 899 |
type.name = "Nail" |
| 900 |
# type = { edible = false } # INVALID |
| 901 |
``` |
| 902 |
|
| 903 |
## Array of Tables |
| 904 |
|
| 905 |
The last syntax that has not yet been described allows writing arrays of tables. |
| 906 |
These can be expressed by using a header with a name in double brackets. The |
| 907 |
first instance of that header defines the array and its first table element, and |
| 908 |
each subsequent instance creates and defines a new table element in that array. |
| 909 |
The tables are inserted into the array in the order encountered. |
| 910 |
|
| 911 |
```toml |
| 912 |
[[product]] |
| 913 |
name = "Hammer" |
| 914 |
sku = 738594937 |
| 915 |
|
| 916 |
[[product]] # empty table within the array |
| 917 |
|
| 918 |
[[product]] |
| 919 |
name = "Nail" |
| 920 |
sku = 284758393 |
| 921 |
|
| 922 |
color = "gray" |
| 923 |
``` |
| 924 |
|
| 925 |
In JSON land, that would give you the following structure. |
| 926 |
|
| 927 |
```json |
| 928 |
{ |
| 929 |
"product": [ |
| 930 |
{ "name": "Hammer", "sku": 738594937 }, |
| 931 |
{}, |
| 932 |
{ "name": "Nail", "sku": 284758393, "color": "gray" } |
| 933 |
] |
| 934 |
} |
| 935 |
``` |
| 936 |
|
| 937 |
Any reference to an array of tables points to the most recently defined table |
| 938 |
element of the array. This allows you to define sub-tables, and even sub-arrays |
| 939 |
of tables, inside the most recent table. |
| 940 |
|
| 941 |
```toml |
| 942 |
[[fruits]] |
| 943 |
name = "apple" |
| 944 |
|
| 945 |
[fruits.physical] # subtable |
| 946 |
color = "red" |
| 947 |
shape = "round" |
| 948 |
|
| 949 |
[[fruits.varieties]] # nested array of tables |
| 950 |
name = "red delicious" |
| 951 |
|
| 952 |
[[fruits.varieties]] |
| 953 |
name = "granny smith" |
| 954 |
|
| 955 |
|
| 956 |
[[fruits]] |
| 957 |
name = "banana" |
| 958 |
|
| 959 |
[[fruits.varieties]] |
| 960 |
name = "plantain" |
| 961 |
``` |
| 962 |
|
| 963 |
The above TOML maps to the following JSON. |
| 964 |
|
| 965 |
```json |
| 966 |
{ |
| 967 |
"fruits": [ |
| 968 |
{ |
| 969 |
"name": "apple", |
| 970 |
"physical": { |
| 971 |
"color": "red", |
| 972 |
"shape": "round" |
| 973 |
}, |
| 974 |
"varieties": [{ "name": "red delicious" }, { "name": "granny smith" }] |
| 975 |
}, |
| 976 |
{ |
| 977 |
"name": "banana", |
| 978 |
"varieties": [{ "name": "plantain" }] |
| 979 |
} |
| 980 |
] |
| 981 |
} |
| 982 |
``` |
| 983 |
|
| 984 |
If the parent of a table or array of tables is an array element, that element |
| 985 |
must already have been defined before the child can be defined. Attempts to |
| 986 |
reverse that ordering must produce an error at parse time. |
| 987 |
|
| 988 |
``` |
| 989 |
# INVALID TOML DOC |
| 990 |
[fruit.physical] # subtable, but to which parent element should it belong? |
| 991 |
color = "red" |
| 992 |
shape = "round" |
| 993 |
|
| 994 |
[[fruit]] # parser must throw an error upon discovering that "fruit" is |
| 995 |
# an array rather than a table |
| 996 |
name = "apple" |
| 997 |
``` |
| 998 |
|
| 999 |
Attempting to append to a statically defined array, even if that array is empty, |
| 1000 |
must produce an error at parse time. |
| 1001 |
|
| 1002 |
``` |
| 1003 |
# INVALID TOML DOC |
| 1004 |
fruits = [] |
| 1005 |
|
| 1006 |
[[fruits]] # Not allowed |
| 1007 |
``` |
| 1008 |
|
| 1009 |
Attempting to define a normal table with the same name as an already established |
| 1010 |
array must produce an error at parse time. Attempting to redefine a normal table |
| 1011 |
as an array must likewise produce a parse-time error. |
| 1012 |
|
| 1013 |
``` |
| 1014 |
# INVALID TOML DOC |
| 1015 |
[[fruits]] |
| 1016 |
name = "apple" |
| 1017 |
|
| 1018 |
[[fruits.varieties]] |
| 1019 |
name = "red delicious" |
| 1020 |
|
| 1021 |
# INVALID: This table conflicts with the previous array of tables |
| 1022 |
[fruits.varieties] |
| 1023 |
name = "granny smith" |
| 1024 |
|
| 1025 |
[fruits.physical] |
| 1026 |
color = "red" |
| 1027 |
shape = "round" |
| 1028 |
|
| 1029 |
# INVALID: This array of tables conflicts with the previous table |
| 1030 |
[[fruits.physical]] |
| 1031 |
color = "green" |
| 1032 |
``` |
| 1033 |
|
| 1034 |
You may also use inline tables where appropriate: |
| 1035 |
|
| 1036 |
```toml |
| 1037 |
points = [ { x = 1, y = 2, z = 3 }, |
| 1038 |
{ x = 7, y = 8, z = 9 }, |
| 1039 |
{ x = 2, y = 4, z = 8 } ] |
| 1040 |
``` |
| 1041 |
|
| 1042 |
## Filename Extension |
| 1043 |
|
| 1044 |
TOML files should use the extension `.toml`. |
| 1045 |
|
| 1046 |
## MIME Type |
| 1047 |
|
| 1048 |
When transferring TOML files over the internet, the appropriate MIME type is |
| 1049 |
`application/toml`. |
| 1050 |
|
| 1051 |
## ABNF Grammar |
| 1052 |
|
| 1053 |
A formal description of TOML's syntax is available, as a separate [ABNF |
| 1054 |
file][abnf]. |
| 1055 |
|
| 1056 |
[abnf]: ./toml.abnf |