|
| 1 | +--- |
| 2 | +title: Language Specification |
| 3 | +slug: adventure/minimessage/specification |
| 4 | +description: A developer-facing specification of the MiniMessage format. |
| 5 | +tableOfContents: |
| 6 | + minHeadingLevel: 2 |
| 7 | + maxHeadingLevel: 5 |
| 8 | +--- |
| 9 | + |
| 10 | +This document outlines the MiniMessage format in detail to aid developers who wish to implement their own MiniMessage |
| 11 | +parser from scratch or understand the internal processes happening during the parsing of MiniMesssage formatted strings. |
| 12 | + |
| 13 | +The keywords “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, |
| 14 | +“MAY”, and “OPTIONAL” in this document are to be interpreted as described in |
| 15 | +[RFC 2119](https://www.rfc-editor.org/rfc/rfc2119.html). |
| 16 | + |
| 17 | +## The MiniMessage language |
| 18 | + |
| 19 | +The MiniMessage language is a markup format used for representing Minecraft's component-based text |
| 20 | +system in a human-readable and modifiable way. Broadly speaking, the language consists of two types |
| 21 | +of tokens: **plain text** and **tags**. |
| 22 | + |
| 23 | +Plain text is any string. This string is UTF-16 compatible. The following is an example of a valid |
| 24 | +plain text part of a MiniMessage-formatted string: |
| 25 | + |
| 26 | +```mm |
| 27 | +The MiniMessage format was made to be as simple as possible. |
| 28 | +Emojies are allowed 😅. So are japanese characters, like 紙. |
| 29 | +``` |
| 30 | + |
| 31 | +MiniMessage tags are primarily used for adding markup information to plain text parts. They can, however, |
| 32 | +also add entirely new content into the serialized component. The way how a tag is resolved makes no |
| 33 | +difference to the MiniMessage lexer. A tag has the following structure: |
| 34 | + |
| 35 | +```mm |
| 36 | +<tagname flags named_argument=value :a sequenced argument:another one> |
| 37 | +``` |
| 38 | + |
| 39 | +A tag consists of the following parts: |
| 40 | + |
| 41 | +- `< >`: All tags are surrounded by less than and more than symbols. |
| 42 | +- `tagname`: Every tag starts with the name. The name follows the list of characters mentioned as allowed |
| 43 | + in the [misc/identifiers](#identifiers) section. |
| 44 | +- Tags can have arguments. There are two distinctions between argument types: named and sequenced. Named arguments |
| 45 | + are, as the name implies, named in some way. Sequenced arguments do not have a name, instead they are a simple list |
| 46 | + of string values. [Tag argument documentation can be found later in the page](#tag-arguments). |
| 47 | + |
| 48 | +### Tag syntax |
| 49 | + |
| 50 | +MiniMessage tags can surround text. |
| 51 | + |
| 52 | +```mm |
| 53 | +<tagname>Inner text</tagname> and outer text. |
| 54 | +``` |
| 55 | + |
| 56 | +Tags can be closed by repeating the tag, with a slash in front of the name. Tags are closed implicitly |
| 57 | +when the end of the string is reached. Furthermore, tags can be nested: |
| 58 | + |
| 59 | +```mm |
| 60 | +<first_tag>Some text <second_tag>even more text</second_tag>, and that's really it! |
| 61 | +``` |
| 62 | + |
| 63 | +Nested tags are closed implicitly when the outer tag is closed. |
| 64 | + |
| 65 | +```mm |
| 66 | +This text is unmarked <outer_tag>marked, <inner_tag>inner</outer_tag>, and again no longer marked. |
| 67 | +``` |
| 68 | + |
| 69 | +If a tag has arguments, these must not be repeated on the closing tag. |
| 70 | + |
| 71 | +```mm |
| 72 | +<tagname:with an argument>Some text</tagname> |
| 73 | +``` |
| 74 | + |
| 75 | +Lastly, normal tags can be closed instantly by prepending a `/` to the more-than symbol of an opening tag. |
| 76 | + |
| 77 | +```mm |
| 78 | +This tag is auto-closed: <tagname/> |
| 79 | +``` |
| 80 | + |
| 81 | +## Tag arguments |
| 82 | + |
| 83 | +Arguments are placed between the tag name and the closing more-than symbol. |
| 84 | + |
| 85 | +```mm |
| 86 | +<tagname [arguments here]> |
| 87 | +``` |
| 88 | + |
| 89 | +### Named argument types |
| 90 | + |
| 91 | +Before each named argument, a piece of [whitespace](#whitespace) must be present. |
| 92 | + |
| 93 | +There exists two types of named arguments: value-based arguments and flag arguments. |
| 94 | + |
| 95 | +#### Flag argument type |
| 96 | + |
| 97 | +Flags may be preceded by a single exclamation mark `!` and must follow the rules set by [identifiers](#identifiers). |
| 98 | + |
| 99 | +```mm |
| 100 | +<tagname some_flag another_flag_4_you !inverted_flag> |
| 101 | +``` |
| 102 | + |
| 103 | +The following shows a tag with invalid flags: |
| 104 | + |
| 105 | +```mm |
| 106 | +<tagname SomeCoolFlag !!double_inverted what-even-is-happening-here?> |
| 107 | +``` |
| 108 | + |
| 109 | +#### Valued argument type |
| 110 | + |
| 111 | +Named arguments with a value consist of an identifier, an equal symbol `=`, and a value. |
| 112 | + |
| 113 | +The identifier follows the rules as explained in the [misc/identifiers](#identifiers) section of this page. |
| 114 | +The value may consist of any UTF-16 characters, but must not contain any whitespace, unless explicitly quoted. |
| 115 | +Please refer to [misc/quoting](#quoting) for any specifics. |
| 116 | + |
| 117 | +Here is an example for valid valued named arguments: |
| 118 | + |
| 119 | +```mm |
| 120 | +<birdtag bird=parrot color='red and blue'> |
| 121 | +``` |
| 122 | + |
| 123 | +And example for invalid valued named arguments: |
| 124 | + |
| 125 | +```mm |
| 126 | +<birbtag vöglein=papagei Colour=red and blue> |
| 127 | +``` |
| 128 | + |
| 129 | +:::note |
| 130 | + |
| 131 | +The above tag, assuming the identifiers were valid, would actually parse both `and` and `blue` as flags. |
| 132 | + |
| 133 | +::: |
| 134 | + |
| 135 | +#### Combining flags and values |
| 136 | + |
| 137 | +These two named types can be combined in any way. |
| 138 | + |
| 139 | +```mm |
| 140 | +<combined_tag aflag some=value !inverted_flag really=yeah!> |
| 141 | +``` |
| 142 | + |
| 143 | +### Sequential arguments |
| 144 | + |
| 145 | +Sequential arguments are declared at the end of the tag. Each sequential argument starts with a colon `:`. |
| 146 | +Unless named arguments are present, a whitespace before the first colon `:` is not necessary. |
| 147 | + |
| 148 | +Sequential arguments may contain any UTF-16 characters. Any instances of `<`, `>`, or `:` characters |
| 149 | +must either be escaped (see [misc/escaping](#escaping)) or the argument must be wrapped in quotes |
| 150 | +(see [misc/quoting](#quoting)). |
| 151 | + |
| 152 | +The following are valid MiniMessage tags with sequential arguments: |
| 153 | + |
| 154 | +```mm |
| 155 | +<simpletag:hey there> |
| 156 | +
|
| 157 | +<another:first argument:second argument> |
| 158 | +
|
| 159 | +<with_whitespace :this is perfectly fine> |
| 160 | +
|
| 161 | +<nested_mm:\<some_cool_tag\> and a \: colon!> |
| 162 | +
|
| 163 | +<nested_mm:"<some_cool_tag> and a : colon, but it's quoted!"> |
| 164 | +``` |
| 165 | + |
| 166 | +### Combining argument types |
| 167 | + |
| 168 | +Named and sequential arguments can be used together. The general syntax looks as follows: |
| 169 | + |
| 170 | +```mm |
| 171 | +<tagname[ named_arguments ][:sequenced:arguments]> |
| 172 | +``` |
| 173 | + |
| 174 | +All named arguments must be located between the tag name and the first non-value colon. |
| 175 | + |
| 176 | +A few examples for valid tags making use of both named and sequenced arguments: |
| 177 | + |
| 178 | +```mm |
| 179 | +<combined coolness=true flags :and sequenced args> |
| 180 | +
|
| 181 | +<combined flags !over more_flags and even !more flagss :I'd call this cool:Would you?:Yeah for sure> |
| 182 | +
|
| 183 | +<combined tic=tac :time's up!> |
| 184 | +``` |
| 185 | + |
| 186 | +## Misc |
| 187 | + |
| 188 | +This section defines miscellaneous behavior of common parts. |
| 189 | + |
| 190 | +### Identifiers |
| 191 | + |
| 192 | +All identifiers must be lowercased and contain only alphanumerical characters or `_`. All identifiers |
| 193 | +used as named argument names should be unique. |
| 194 | + |
| 195 | +### Quoting |
| 196 | +Argument values can be quoted. A value counts as quoted if the first character is a `'` or `"`. The quoted |
| 197 | +value ends as soon as another unquoted quote of the same character as the starting quote is found at the |
| 198 | +end of an argument. |
| 199 | + |
| 200 | +Between the opening and the closing quote, any UTF-16 characters may be present. This also includes the same |
| 201 | +quote as used for the string. The following would be a valid tag; |
| 202 | + |
| 203 | +```mm |
| 204 | +<tag:"double quoted", yet contains a double quote?"> |
| 205 | +``` |
| 206 | + |
| 207 | +This is because the `"` in the middle is **not the last character of the value**. Therefore, it is read |
| 208 | +literally, since the tag would otherwise be invalid. |
| 209 | + |
| 210 | +:::tip |
| 211 | + |
| 212 | +As long as the quote is not closed, the lexer must continue reading characters. If the end of the |
| 213 | +input is reached before closing quote is found, the tag and any following characters should be |
| 214 | +read as plain text, as the tag is never closed. This is to aid users in finding the error in their syntax. |
| 215 | + |
| 216 | +::: |
| 217 | + |
| 218 | +### Whitespace |
| 219 | + |
| 220 | +A whitespace character may be a classical space `\s`, a tab character `\t`, |
| 221 | +a newline `\n`, or a carriage return `\r`. |
| 222 | + |
| 223 | +### Escaping |
| 224 | + |
| 225 | +In MiniMessage, certain symbols, which would be interpreted differently by a lexer may be preceded by a backslash `\` |
| 226 | +to instead be included literally. This includes backslash `\` characters, if they would have any effect on the next |
| 227 | +symbol. If a backlash character had no effect, it is included literally. |
| 228 | + |
| 229 | +## Formal grammar |
| 230 | + |
| 231 | +This segment declares the formal grammar (in a flavor of the Backus-Naur form) which specifies the MiniMessage language. |
| 232 | + |
| 233 | +The specific flavor used here changes that non-terminal symbols are no longer enclosed in angle brackets `<>` |
| 234 | +and the `::=` meta symbol is replaced by `→`. Curly brackets `{}` declare optional parts. Lastly, a `+` suffix |
| 235 | +declares that a symbol should appear at least once, but may appear more often, whilst a `*` suffix declares that |
| 236 | +a symbol may appear once or more often. |
| 237 | + |
| 238 | +```bnf |
| 239 | +; Important notes regarding this specific grammar: due to the massive number of characters included |
| 240 | +; in the UTF-16 characterset, some special non-terminal symbols have been added: |
| 241 | +; |
| 242 | +; utf-16-char → includes all UTF-16 characters. |
| 243 | +; |
| 244 | +; utf-16-char-no-whitespace → includes all UTF-16 characters except for spaces (\s), tabs (\t), newlines (\n) |
| 245 | +; and carriage returns (\r). |
| 246 | +; |
| 247 | +; utf-16-char-no-angle-or-colon → includes all UTF-16 characters except for the |
| 248 | +; angle-bracket characters (<>) and colon (:). However |
| 249 | +; those characters are valid if an uneven number of backslash |
| 250 | +; characters is located infront of them. |
| 251 | +
|
| 252 | +minimessage → string {tag string} |
| 253 | +
|
| 254 | +string → utf-16-char* |
| 255 | +
|
| 256 | +tag → "<" tag-name tag-arguments "/>" |
| 257 | +tag → "<" tag-name tag-arguments ">" minimessage {"</" tag-name ">"} |
| 258 | +
|
| 259 | +tag-name → identifier |
| 260 | +
|
| 261 | +tag-arguments → "" | named-argument " "+ sequential-argument | named-argument | " "* sequential-argument |
| 262 | +
|
| 263 | +named-argument → "" | " "+ {"!"} identifier {named-argument} | " "+ identifier "=" named-value {named-argument} |
| 264 | +
|
| 265 | +named-value → "" | quoted | no-whitespace-string |
| 266 | +
|
| 267 | +no-whitespace-string → utf-16-char-no-whitespace* |
| 268 | +
|
| 269 | +sequential-argument → ":" sequential-value {sequential-argument} |
| 270 | +
|
| 271 | +sequential-value → "" | quoted | sequential-string |
| 272 | +
|
| 273 | +sequential-string → utf-16-char-no-angle-or-colon* |
| 274 | +
|
| 275 | +quoted → "'" string "'" | """ string """ |
| 276 | +
|
| 277 | +identifier → alphanumeric+ |
| 278 | +
|
| 279 | +alphanumeric → "a" | "b" | "c" | "d" |
| 280 | + | "e" | "f" | "g" | "h" |
| 281 | + | "i" | "j" | "k" | "l" |
| 282 | + | "m" | "n" | "o" | "p" |
| 283 | + | "q" | "r" | "s" | "t" |
| 284 | + | "u" | "v" | "w" | "x" |
| 285 | + | "y" | "z" | "_" | "0" |
| 286 | + | "1" | "2" | "3" | "4" |
| 287 | + | "5" | "6" | "7" | "8" |
| 288 | + | "9" |
| 289 | +``` |
0 commit comments