1 2# Oniguruma syntax (operator) configuration 3 4_Documented for Oniguruma 6.9.3 (2019/08/08)_ 5 6 7---------- 8 9 10## Overview 11 12This document details how to configure Oniguruma's syntax, by describing the desired 13syntax operators and behaviors in an instance of the OnigSyntaxType struct, just like 14the built-in Oniguruma syntaxes do. 15 16Configuration operators are bit flags, and are broken into multiple groups, somewhat arbitrarily, 17because Oniguruma takes its configuration as a trio of 32-bit `unsigned int` values, assigned as 18the first three fields in an `OnigSyntaxType` struct: 19 20```C 21typedef struct { 22 unsigned int op; 23 unsigned int op2; 24 unsigned int behavior; 25 OnigOptionType options; /* default option */ 26 OnigMetaCharTableType meta_char_table; 27} OnigSyntaxType; 28``` 29 30The first group of configuration flags (`op`) roughly corresponds to the 31configuration for "basic regex." The second group (`op2`) roughly corresponds 32to the configuration for "advanced regex." And the third group (`behavior`) 33describes more-or-less what to do for broken input, bad input, or other corner-case 34regular expressions whose meaning is not well-defined. These three groups of 35flags are described in full below, and tables of their usages for various syntaxes 36follow. 37 38The `options` field describes the default compile options to use if the caller does 39not specify any options when invoking `onig_new()`. 40 41The `meta_char_table` field is used exclusively by the ONIG_SYN_OP_VARIABLE_META_CHARACTERS 42option, which allows the various regex metacharacters, like `*` and `?`, to be replaced 43with alternates (for example, SQL typically uses `%` instead of `.*` and `_` instead of `?`). 44 45 46---------- 47 48 49## Group One Flags (op) 50 51 52This group contains "basic regex" constructs, features common to most regex systems. 53 54 55### 0. ONIG_SYN_OP_VARIABLE_META_CHARACTERS 56 57_Set in: none_ 58 59Enables support for `onig_set_meta_char()`, which allows you to provide alternate 60characters that will be used instead of the six special characters that are normally 61these characters below: 62 63 - `ONIG_META_CHAR_ESCAPE`: `\` 64 - `ONIG_META_CHAR_ANYCHAR`: `.` 65 - `ONIG_META_CHAR_ANYTIME`: `*` 66 - `ONIG_META_CHAR_ZERO_OR_ONE_TIME`: `?` 67 - `ONIG_META_CHAR_ONE_OR_MORE_TIME`: `+` 68 - `ONIG_META_CHAR_ANYCHAR_ANYTIME`: Equivalent in normal regex to `.*`, but supported 69 explicitly so that Oniguruma can support matching SQL `%` wildcards or shell `*` wildcards. 70 71If this flag is set, then the values defined using `onig_set_meta_char()` will be used; 72if this flag is clear, then the default regex characters will be used instead, and 73data set by `onig_set_meta_char()` will be ignored. 74 75 76### 1. ONIG_SYN_OP_DOT_ANYCHAR (enable `.`) 77 78_Set in: PosixBasic, PosixExtended, Emacs, Grep, GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 79 80Enables support for the standard `.` metacharacter, meaning "any one character." You 81usually want this flag on unless you have turned on `ONIG_SYN_OP_VARIABLE_META_CHARACTERS` 82so that you can use a metacharacter other than `.` instead. 83 84 85### 2. ONIG_SYN_OP_ASTERISK_ZERO_INF (enable `r*`) 86 87_Set in: PosixBasic, PosixExtended, Emacs, Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 88 89Enables support for the standard `r*` metacharacter, meaning "zero or more r's." 90You usually want this flag set unless you have turned on `ONIG_SYN_OP_VARIABLE_META_CHARACTERS` 91so that you can use a metacharacter other than `*` instead. 92 93 94### 3. ONIG_SYN_OP_ESC_ASTERISK_ZERO_INF (enable `r\*`) 95 96_Set in: none_ 97 98Enables support for an escaped `r\*` metacharacter, meaning "zero or more r's." This is 99useful if you have disabled support for the normal `r*` metacharacter because you want `*` 100to simply match a literal `*` character, but you still want some way of activating "zero or more" 101behavior. 102 103 104### 4. ONIG_SYN_OP_PLUS_ONE_INF (enable `r+`) 105 106_Set in: PosixExtended, Emacs, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 107 108Enables support for the standard `r+` metacharacter, meaning "one or more r's." 109You usually want this flag set unless you have turned on `ONIG_SYN_OP_VARIABLE_META_CHARACTERS` 110so that you can use a metacharacter other than `+` instead. 111 112 113### 5. ONIG_SYN_OP_ESC_PLUS_ONE_INF (enable `r\+`) 114 115_Set in: Grep_ 116 117Enables support for an escaped `r\+` metacharacter, meaning "one or more r's." This is 118useful if you have disabled support for the normal `r+` metacharacter because you want `+` 119to simply match a literal `+` character, but you still want some way of activating "one or more" 120behavior. 121 122 123### 6. ONIG_SYN_OP_QMARK_ZERO_ONE (enable `r?`) 124 125_Set in: PosixExtended, Emacs, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 126 127Enables support for the standard `r?` metacharacter, meaning "zero or one r" or "an optional r." 128You usually want this flag set unless you have turned on `ONIG_SYN_OP_VARIABLE_META_CHARACTERS` 129so that you can use a metacharacter other than `?` instead. 130 131 132### 7. ONIG_SYN_OP_ESC_QMARK_ZERO_ONE (enable `r\?`) 133 134_Set in: Grep_ 135 136Enables support for an escaped `r\?` metacharacter, meaning "zero or one r" or "an optional 137r." This is useful if you have disabled support for the normal `r?` metacharacter because 138you want `?` to simply match a literal `?` character, but you still want some way of activating 139"optional" behavior. 140 141 142### 8. ONIG_SYN_OP_BRACE_INTERVAL (enable `r{l,u}`) 143 144_Set in: PosixExtended, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 145 146Enables support for the `r{lower,upper}` range form, common to more advanced 147regex engines, which lets you specify precisely a minimum and maximum range on how many r's 148must match (and not simply "zero or more"). 149 150This form also allows `r{count}` to specify a precise count of r's that must match. 151 152This form also allows `r{lower,}` to be equivalent to `r{lower,infinity}`. 153 154If and only if the `ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV` behavior flag is set, 155this form also allows `r{,upper}` to be equivalent to `r{0,upper}`; otherwise, 156`r{,upper}` will be treated as an error. 157 158 159### 9. ONIG_SYN_OP_ESC_BRACE_INTERVAL (enable `\{` and `\}`) 160 161_Set in: PosixBasic, Emacs, Grep_ 162 163Enables support for an escaped `r\{lower,upper\}` range form. This is useful if you 164have disabled support for the normal `r{...}` range form and want curly braces to simply 165match literal curly brace characters, but you still want some way of activating 166"range" behavior. 167 168 169### 10. ONIG_SYN_OP_VBAR_ALT (enable `r|s`) 170 171_Set in: PosixExtended, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 172 173Enables support for the common `r|s` alternation operator. You usually want this 174flag set. 175 176 177### 11. ONIG_SYN_OP_ESC_VBAR_ALT (enable `\|`) 178 179_Set in: Emacs, Grep_ 180 181Enables support for an escaped `r\|s` alternation form. This is useful if you 182have disabled support for the normal `r|s` alternation form and want `|` to simply 183match a literal `|` character, but you still want some way of activating "alternate" behavior. 184 185 186### 12. ONIG_SYN_OP_LPAREN_SUBEXP (enable `(r)`) 187 188_Set in: PosixExtended, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 189 190Enables support for the common `(...)` grouping-and-capturing operators. You usually 191want this flag set. 192 193 194### 13. ONIG_SYN_OP_ESC_LPAREN_SUBEXP (enable `\(` and `\)`) 195 196_Set in: PosixBasic, Emacs, Grep_ 197 198Enables support for escaped `\(...\)` grouping-and-capturing operators. This is useful if you 199have disabled support for the normal `(...)` grouping-and-capturing operators and want 200parentheses to simply match literal parenthesis characters, but you still want some way of 201activating "grouping" or "capturing" behavior. 202 203 204### 14. ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR (enable `\A` and `\Z` and `\z`) 205 206_Set in: GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 207 208Enables support for the anchors `\A` (start-of-string), `\Z` (end-of-string or 209newline-at-end-of-string), and `\z` (end-of-string) escapes. 210 211(If the escape metacharacter has been changed from the default of `\`, this 212option will recognize that metacharacter instead.) 213 214 215### 15. ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR (enable `\G`) 216 217_Set in: GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 218 219Enables support for the special anchor `\G` (start-of-previous-match). 220 221(If the escape metacharacter has been changed from the default of `\`, this 222option will recognize that metacharacter instead.) 223 224Note that `OnigRegex`/`regex_t` are not stateful objects, and do _not_ record 225the location of the previous match. The `\G` flag uses the `start` parameter 226explicitly passed to `onig_search()` (or `onig_search_with_param()` to determine 227the "start of the previous match," so if the caller always passes the start of 228the entire buffer as the function's `start` parameter, then `\G` will behave 229exactly the same as `\A`. 230 231 232### 16. ONIG_SYN_OP_DECIMAL_BACKREF (enable `\num`) 233 234_Set in: PosixBasic, PosixExtended, Emacs, Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 235 236Enables support for subsequent matches to back references to prior capture groups `(...)` using 237the common `\num` syntax (like `\3`). 238 239If this flag is clear, then a numeric escape like `\3` will either be treated as a literal `3`, 240or, if `ONIG_SYN_OP_ESC_OCTAL3` is set, will be treated as an octal character code `\3`. 241 242You usually want this enabled, and it is enabled by default in every built-in syntax. 243 244 245### 17. ONIG_SYN_OP_BRACKET_CC (enable `[...]`) 246 247_Set in: PosixBasic, PosixExtended, Emacs, Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 248 249Enables support for recognizing character classes, like `[a-z]`. If this flag is not set, `[` 250and `]` will be treated as ordinary literal characters instead of as metacharacters. 251 252You usually want this enabled, and it is enabled by default in every built-in syntax. 253 254 255### 18. ONIG_SYN_OP_ESC_W_WORD (enable `\w` and `\W`) 256 257_Set in: Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 258 259Enables support for the common `\w` and `\W` shorthand forms. These match "word characters," 260whose meaning varies depending on the encoding being used. 261 262In ASCII encoding, `\w` is equivalent to `[A-Za-z0-9_]`. 263 264In most other encodings, `\w` matches many more characters, including accented letters, Greek letters, 265Cyrillic letters, Braille letters and numbers, Runic letters, Hebrew letters, Arabic letters and numerals, 266Chinese Han ideographs, Japanese Katakana and Hiragana, Korean Hangul, and generally any symbol that 267could qualify as a phonetic "letter" or counting "number" in any language. (Note that emoji are _not_ 268considered "word characters.") 269 270`\W` always matches the opposite of whatever `\w` matches. 271 272 273### 19. ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END (enable `\<` and `\>`) 274 275_Set in: Grep, GnuRegex_ 276 277Enables support for the GNU-specific `\<` and `\>` word-boundary metacharacters. These work like 278the `\b` word-boundary metacharacter, but only match at one end of the word or the other: `\<` 279only matches at a transition from a non-word character to a word character (i.e., at the start 280of a word), and `\>` only matches at a transition from a word character to a non-word character 281(i.e., at the end of a word). 282 283Most regex syntaxes do _not_ support these metacharacters. 284 285 286### 20. ONIG_SYN_OP_ESC_B_WORD_BOUND (enable `\b` and `\B`) 287 288_Set in: Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 289 290Enables support for the common `\b` and `\B` word-boundary metacharacters. The `\b` metacharacter 291matches a zero-width position at a transition from word-characters to non-word-characters, or vice 292versa. The `\B` metacharacter matches at all positions _not_ matched by `\b`. 293 294See details in `ONIG_SYN_OP_ESC_W_WORD` above for an explanation as to which characters 295are considered "word characters." 296 297 298### 21. ONIG_SYN_OP_ESC_S_WHITE_SPACE (enable `\s` and `\S`) 299 300_Set in: GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 301 302Enables support for the common `\s` and `\S` whitespace-matching metacharacters. 303 304The `\s` metacharacter in ASCII encoding is exactly equivalent to the character class 305`[\t\n\v\f\r ]`, or characters codes 9 through 13 (inclusive), and 32. 306 307The `\s` metacharacter in Unicode is exactly equivalent to the character class 308`[\t\n\v\f\r \x85\xA0\x1680\x2000-\x200A\x2028-\x2029\x202F\x205F\x3000]` — that is, it matches 309the same as ASCII, plus U+0085 (next line), U+00A0 (nonbreaking space), U+1680 (Ogham space mark), 310U+2000 (en quad) through U+200A (hair space) (this range includes several widths of Unicode spaces), 311U+2028 (line separator) through U+2029 (paragraph separator), 312U+202F (narrow no-break space), U+205F (medium mathematical space), and U+3000 (CJK ideographic space). 313 314All non-Unicode encodings are handled by converting their code points to the appropriate 315Unicode-equivalent code points, and then matching according to Unicode rules. 316 317`\S` always matches any one character that is _not_ in the set matched by `\s`. 318 319 320### 22. ONIG_SYN_OP_ESC_D_DIGIT (enable `\d` and `\D`) 321 322_Set in: GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 323 324Enables support for the common `\d` and `\D` digit-matching metacharacters. 325 326The `\d` metacharacter in ASCII encoding is exactly equivalent to the character class 327`[0-9]`, or characters codes 48 through 57 (inclusive). 328 329The `\d` metacharacter in Unicode matches `[0-9]`, as well as digits in Arabic, Devanagari, 330Bengali, Laotian, Mongolian, CJK fullwidth numerals, and many more. 331 332All non-Unicode encodings are handled by converting their code points to the appropriate 333Unicode-equivalent code points, and then matching according to Unicode rules. 334 335`\D` always matches any one character that is _not_ in the set matched by `\d`. 336 337 338### 23. ONIG_SYN_OP_LINE_ANCHOR (enable `^r` and `r$`) 339 340_Set in: Emacs, Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 341 342Enables support for the common `^` and `$` line-anchor metacharacters. 343 344In single-line mode, `^` matches the start of the input buffer, and `$` matches 345the end of the input buffer. In multi-line mode, `^` matches if the preceding 346character is `\n`; and `$` matches if the following character is `\n`. 347 348(Note that Oniguruma does not recognize other newline types: It only matches 349`^` and `$` against `\n`: not `\r`, not `\r\n`, not the U+2028 line separator, 350and not any other form.) 351 352 353### 24. ONIG_SYN_OP_POSIX_BRACKET (enable POSIX `[:xxxx:]`) 354 355_Set in: PosixBasic, PosixExtended, Grep, GnuRegex, Perl, Java, Perl_NG, Ruby, Oniguruma_ 356 357Enables support for the POSIX `[:xxxx:]` character classes, like `[:alpha:]` and `[:digit:]`. 358The supported POSIX character classes are `alnum`, `alpha`, `blank`, `cntrl`, `digit`, 359`graph`, `lower`, `print`, `punct`, `space`, `upper`, `xdigit`, `ascii`, `word`. 360 361 362### 25. ONIG_SYN_OP_QMARK_NON_GREEDY (enable `r??`, `r*?`, `r+?`, and `r{n,m}?`) 363 364_Set in: Perl, Java, Perl_NG, Ruby, Oniguruma_ 365 366Enables support for lazy (non-greedy) quantifiers: That is, if you append a `?` after 367another quantifier such as `?`, `*`, `+`, or `{n,m}`, Oniguruma will try to match 368as _little_ as possible instead of as _much_ as possible. 369 370 371### 26. ONIG_SYN_OP_ESC_CONTROL_CHARS (enable `\n`, `\r`, `\t`, etc.) 372 373_Set in: PosixBasic, PosixExtended, Java, Perl, Perl_NG, Ruby, Oniguruma_ 374 375Enables support for C-style control-code escapes, like `\n` and `\r`. Specifically, 376this recognizes `\a` (7), `\b` (8), `\t` (9), `\n` (10), `\f` (12), `\r` (13), and 377`\e` (27). If ONIG_SYN_OP2_ESC_V_VTAB is enabled (see below), this also enables 378support for recognizing `\v` as code point 11. 379 380 381### 27. ONIG_SYN_OP_ESC_C_CONTROL (enable `\cx` control codes) 382 383_Set in: Java, Perl, Perl_NG, Ruby, Oniguruma_ 384 385Enables support for named control-code escapes, like `\cm` or `\cM` for code-point 38613. In this shorthand form, control codes may be specified by `\c` (for "Control") 387followed by an alphabetic letter, a-z or A-Z, indicating which code point to represent 388(1 through 26). So `\cA` is code point 1, and `\cZ` is code point 26. 389 390 391### 28. ONIG_SYN_OP_ESC_OCTAL3 (enable `\OOO` octal codes) 392 393_Set in: Java, Perl, Perl_NG, Ruby, Oniguruma_ 394 395Enables support for octal-style escapes of up to three digits, like `\1` for code 396point 1, and `\177` for code point 127. Octal values greater than 255 will result 397in an error message. 398 399 400### 29. ONIG_SYN_OP_ESC_X_HEX2 (enable `\xHH` hex codes) 401 402_Set in: Java, Perl, Perl_NG, Ruby, Oniguruma_ 403 404Enables support for hexadecimal-style escapes of up to two digits, like `\x1` for code 405point 1, and `\x7F` for code point 127. 406 407 408### 30. ONIG_SYN_OP_ESC_X_BRACE_HEX8 (enable `\x{7HHHHHHH}` hex codes) 409 410_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 411 412Enables support for brace-wrapped hexadecimal-style escapes of up to eight digits, 413like `\x{1}` for code point 1, and `\x{FFFE}` for code point 65534. 414 415 416### 31. ONIG_SYN_OP_ESC_O_BRACE_OCTAL (enable `\o{1OOOOOOOOOO}` octal codes) 417 418_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 419 420Enables support for brace-wrapped octal-style escapes of up to eleven digits, 421like `\o{1}` for code point 1, and `\o{177776}` for code point 65534. 422 423(New feature as of Oniguruma 6.3.) 424 425 426---------- 427 428 429## Group Two Flags (op2) 430 431 432This group contains support for lesser-known regex syntax constructs. 433 434 435### 0. ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE (enable `\Q...\E`) 436 437_Set in: Java, Perl, Perl_NG_ 438 439Enables support for "quoted" parts of a pattern: Between `\Q` and `\E`, all 440syntax parsing is turned off, so that metacharacters like `*` and `+` will no 441longer be treated as metacharacters, and instead will be matched as literal 442`*` and `+`, as if they had been escaped with `\*` and `\+`. 443 444 445### 1. ONIG_SYN_OP2_QMARK_GROUP_EFFECT (enable `(?...)`) 446 447_Set in: Java, Perl, Perl_NG, Ruby, Oniguruma_ 448 449Enables support for the fairly-common `(?...)` grouping operator, which 450controls precedence but which does _not_ capture its contents. 451 452 453### 2. ONIG_SYN_OP2_OPTION_PERL (enable options `(?imsx)` and `(?-imsx)`) 454 455_Set in: Java, Perl, Perl_NG_ 456 457Enables support of regex options. (i,m,s,x) 458The supported toggle-able options for this flag are: 459 460 - `i` - Case-insensitivity 461 - `m` - Multi-line mode (`^` and `$` match at `\n` as well as start/end of buffer) 462 - `s` - Single-line mode (`.` can match `\n`) 463 - `x` - Extended pattern (free-formatting: whitespace will ignored) 464 465 466### 3. ONIG_SYN_OP2_OPTION_RUBY (enable options `(?imx)` and `(?-imx)`) 467 468_Set in: Ruby, Oniguruma_ 469 470Enables support of regex options. (i,m,x) 471The supported toggle-able options for this flag are: 472 473 - `i` - Case-insensitivity 474 - `m` - Multi-line mode (`.` can match `\n`) 475 - `x` - Extended pattern (free-formatting: whitespace will ignored) 476 477 478### 4. ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT (enable `r?+`, `r*+`, and `r++`) 479 480_Set in: Ruby, Oniguruma_ 481 482Enables support for the _possessive_ quantifiers `?+`, `*+`, and `++`, which 483work similarly to `?` and `*` and `+`, respectively, but which do not backtrack 484after matching: Like the normal greedy quantifiers, they match as much as 485possible, but they do not attempt to match _less_ than their maximum possible 486extent if subsequent parts of the pattern fail to match. 487 488 489### 5. ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL (enable `r{n,m}+`) 490 491_Set in: Java_ 492 493Enables support for the _possessive_ quantifier `{n,m}+`, which 494works similarly to `{n,m}`, but which does not backtrack 495after matching: Like the normal greedy quantifier, it matches as much as 496possible, but it do not attempt to match _less_ than its maximum possible 497extent if subsequent parts of the pattern fail to match. 498 499 500### 6. ONIG_SYN_OP2_CCLASS_SET_OP (enable `&&` within `[...]`) 501 502_Set in: Java, Ruby, Oniguruma_ 503 504Enables support for character-class _intersection_. For example, with this 505feature enabled, you can write `[a-z&&[^aeiou]]` to produce a character class 506of only consonants, or `[\0-\37&&[^\n\r]]` to produce a character class of 507all control codes _except_ newlines. 508 509 510### 7. ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP (enable named captures `(?<name>...)`) 511 512_Set in: Perl_NG, Ruby, Oniguruma_ 513 514Enables support for _naming_ capture groups, so that instead of having to 515refer to captures by position (like `\3` or `$3`), you can refer to them by names 516(like `server` and `path`). This supports the Perl/Ruby naming syntaxes `(?<name>...)` 517and `(?'name'...)`, but not the Python `(?P<name>...)` syntax. 518 519 520### 8. ONIG_SYN_OP2_ESC_K_NAMED_BACKREF (enable named backreferences `\k<name>`) 521 522_Set in: Perl_NG, Ruby, Oniguruma_ 523 524Enables support for substituted backreferences by name, not just by position. 525This supports using `\k'name'` in addition to supporting `\k<name>`. This also 526supports an Oniguruma-specific extension that lets you specify the _distance_ of 527the match, if the capture matched multiple times, by writing `\k<name+n>` or 528`\k<name-n>`. 529 530 531### 9. ONIG_SYN_OP2_ESC_G_SUBEXP_CALL (enable backreferences `\g<name>` and `\g<n>`) 532 533_Set in: Perl_NG, Ruby, Oniguruma_ 534 535Enables support for substituted backreferences by both name and position using 536the same syntax. This supports using `\g'name'` and `\g'1'` in addition to 537supporting `\g<name>` and `\g<1>`. 538 539 540### 10. ONIG_SYN_OP2_ATMARK_CAPTURE_HISTORY (enable `(?@...)` and `(?@<name>...)`) 541 542_Set in: none_ 543 544Enables support for _capture history_, which can answer via the `onig_*capture*()` 545functions exactly which captures were matched, how many times, and where in the 546input they were matched, by placing `?@` in front of the capture. Per Oniguruma's 547regex syntax documentation (appendix A-5): 548 549`/(?@a)*/.match("aaa")` ==> `[<0-1>, <1-2>, <2-3>]` 550 551This can require substantial memory, is primarily useful for debugging, and is not 552enabled by default in any syntax. 553 554 555### 11. ONIG_SYN_OP2_ESC_CAPITAL_C_BAR_CONTROL (enable `\C-x`) 556 557_Set in: Ruby, Oniguruma_ 558 559Enables support for Ruby legacy control-code escapes, like `\C-m` or `\C-M` for code-point 56013. In this shorthand form, control codes may be specified by `\C-` (for "Control") 561followed by a single character (or equivalent), indicating which code point to represent, 562based on that character's lowest five bits. So, like `\c`, you can represent code-point 56310 with `\C-j`, but you can also represent it with `\C-*` as well. 564 565See also ONIG_SYN_OP_ESC_C_CONTROL, which enables the more-common `\cx` syntax. 566 567 568### 12. ONIG_SYN_OP2_ESC_CAPITAL_M_BAR_META (enable `\M-x`) 569 570_Set in: Ruby, Oniguruma_ 571 572Enables support for Ruby legacy meta-code escapes. When you write `\M-x`, Oniguruma 573will match an `x` whose 8th bit is set (i.e., the character code of `x` will be or'ed 574with `0x80`). So, for example, you can match `\x81` using `\x81`, or you can write 575`\M-\1`. This is mostly useful when working with legacy 8-bit character encodings. 576 577 578### 13. ONIG_SYN_OP2_ESC_V_VTAB (enable `\v` as vertical tab) 579 580_Set in: Java, Ruby, Oniguruma_ 581 582Enables support for a C-style `\v` escape code, meaning "vertical tab." If enabled, 583`\v` will be equivalent to ASCII code point 11. 584 585 586### 14. ONIG_SYN_OP2_ESC_U_HEX4 (enable `\uHHHH` for Unicode) 587 588_Set in: Java, Ruby, Oniguruma_ 589 590Enables support for a Java-style `\uHHHH` escape code for representing Unicode 591code-points by number, using up to four hexadecimal digits (up to `\uFFFF`). So, 592for example, `\u221E` will match an infinity symbol, `∞`. 593 594For code points larger than four digits, like the emoji `��` (aerial tramway, or code 595point U+1F6A1), you must either represent the character directly using an encoding like 596UTF-8, or you must enable support for ONIG_SYN_OP_ESC_X_BRACE_HEX8 or 597ONIG_SYN_OP_ESC_O_BRACE_OCTAL, which support more than four digits. 598 599(New feature as of Oniguruma 6.7.) 600 601 602### 15. ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR (enable ``\` `` and `\'` anchors) 603 604_Set in: Emacs_ 605 606This flag makes the ``\` `` and `\'` escapes function identically to 607`\A` and `\z`, respectively (when ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR is enabled). 608 609These anchor forms are very obscure, and rarely supported by other regex libraries. 610 611 612### 16. ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY (enable `\p{...}` and `\P{...}`) 613 614_Set in: Java, Perl, Perl_NG, Ruby, Oniguruma_ 615 616Enables support for an alternate syntax for POSIX character classes; instead of 617writing `[:alpha:]` when this is enabled, you can instead write `\p{alpha}`. 618 619See also ONIG_SYN_OP_POSIX_BRACKET for the classic POSIX form. 620 621 622### 17. ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT (enable `\p{^...}` and `\P{^...}`) 623 624_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 625 626Enables support for an alternate syntax for POSIX character classes; instead of 627writing `[:^alpha:]` when this is enabled, you can instead write `\p{^alpha}`. 628 629See also ONIG_SYN_OP_POSIX_BRACKET for the classic POSIX form. 630 631 632### 18. ONIG_SYN_OP2_CHAR_PROPERTY_PREFIX_IS 633 634_(not presently used)_ 635 636 637### 19. ONIG_SYN_OP2_ESC_H_XDIGIT (enable `\h` and `\H`) 638 639_Set in: Ruby, Oniguruma_ 640 641Enables support for the Ruby-specific shorthand `\h` and `\H` metacharacters. 642Somewhat like `\d` matches decimal digits, `\h` matches hexadecimal digits — that is, 643characters in `[0-9a-fA-F]`. 644 645`\H` matches the opposite of whatever `\h` matches. 646 647 648### 20. ONIG_SYN_OP2_INEFFECTIVE_ESCAPE (disable `\`) 649 650_Set in: As-is_ 651 652If set, this disables all escape codes, shorthands, and metacharacters that start 653with `\` (or whatever the configured escape character is), allowing `\` to be treated 654as a literal `\`. 655 656You usually do not want this flag to be enabled. 657 658 659### 21. ONIG_SYN_OP2_QMARK_LPAREN_IF_ELSE (enable `(?(...)then|else)`) 660 661_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 662 663Enables support for conditional inclusion of subsequent regex patterns based on whether 664a prior named or numbered capture matched, or based on whether a pattern will 665match. This supports many different forms, including: 666 667 - `(?(<foo>)then|else)` - condition based on a capture by name. 668 - `(?('foo')then|else)` - condition based on a capture by name. 669 - `(?(3)then|else)` - condition based on a capture by number. 670 - `(?(+3)then|else)` - forward conditional to a future match, by relative position. 671 - `(?(-3)then|else)` - backward conditional to a prior match, by relative position. 672 - `(?(foo)then|else)` - this matches a pattern `foo`. (foo is any sub-expression) 673 674(New feature as of Oniguruma 6.5.) 675 676 677### 22. ONIG_SYN_OP2_ESC_CAPITAL_K_KEEP (enable `\K`) 678 679_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 680 681Enables support for `\K`, which excludes all content before it from the overall 682regex match (i.e., capture #0). So, for example, pattern `foo\Kbar` would match 683`foobar`, but capture #0 would only include `bar`. 684 685(New feature as of Oniguruma 6.5.) 686 687 688### 23. ONIG_SYN_OP2_ESC_CAPITAL_R_GENERAL_NEWLINE (enable `\R`) 689 690_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 691 692Enables support for `\R`, the "general newline" shorthand, which matches 693`(\r\n|[\n\v\f\r\u0085\u2028\u2029])` (obviously, the Unicode values are cannot be 694matched in ASCII encodings). 695 696(New feature as of Oniguruma 6.5.) 697 698 699### 24. ONIG_SYN_OP2_ESC_CAPITAL_N_O_SUPER_DOT (enable `\N` and `\O`) 700 701_Set in: Perl, Perl_NG, Oniguruma_ 702 703Enables support for `\N` and `\O`. `\N` is "not a line break," which is much 704like the standard `.` metacharacter, except that while `.` can be affected by 705the single-line setting, `\N` always matches exactly one character that is not 706one of the various line-break characters (like `\n` and `\r`). 707 708`\O` matches exactly one character, regardless of whether single-line or 709multi-line mode are enabled or disabled. 710 711(New feature as of Oniguruma 6.5.) 712 713 714### 25. ONIG_SYN_OP2_QMARK_TILDE_ABSENT_GROUP (enable `(?~...)`) 715 716_Set in: Ruby, Oniguruma_ 717 718Enables support for the `(?~r)` "absent operator" syntax, which matches 719as much as possible as long as the result _doesn't_ match pattern `r`. This is 720_not_ the same as negative lookahead or negative lookbehind. 721 722Among the most useful examples of this is `\/\*(?~\*\/)\*\/`, which matches 723C-style comments by simply saying "starts with /*, ends with */, and _doesn't_ 724contain a */ in between." 725 726A full explanation of this feature is complicated, but it is useful, and an 727excellent article about it is [available on Medium](https://medium.com/rubyinside/the-new-absent-operator-in-ruby-s-regular-expressions-7c3ef6cd0b99). 728 729(New feature as of Oniguruma 6.5.) 730 731 732### 26. ONIG_SYN_OP2_ESC_X_Y_TEXT_SEGMENT (enable `\X` and `\Y` and `\y`) 733 734_Set in: Perl, Perl_NG, Ruby, Oniguruma_ 735 736`\X` is another variation on `.`, designed to support Unicode, in that it matches 737a full _grapheme cluster_. In Unicode, `à` can be encoded as one code point, 738`U+00E0`, or as two, `U+0061 U+0300`. If those are further escaped using UTF-8, 739the former becomes two bytes, and the latter becomes three. Unfortunately, `.` 740would naively match only one or two bytes, depending on the encoding, and would 741likely incorrectly match anything from just `a` to a broken half of a code point. 742`\X` is designed to fix this: It matches the full `à`, no matter how `à` is 743encoded or decomposed. 744 745`\y` matches a cluster boundary, i.e., a zero-width position between 746graphemes, somewhat like `\b` matches boundaries between words. `\Y` matches 747the _opposite_ of `\y`, that is, a zero-width position between code points in 748the _middle_ of a grapheme. 749 750(New feature as of Oniguruma 6.6.) 751 752 753### 27. ONIG_SYN_OP2_QMARK_PERL_SUBEXP_CALL (enable `(?R)` and `(?&name)`) 754 755_Set in: Perl_NG_ 756 757Enables support for substituted backreferences by both name and position using 758Perl-5-specific syntax. This supports using `(?R3)` and `(?&name)` to reference 759previous (and future) matches, similar to the more-common `\g<3>` and `\g<name>` 760backreferences. 761 762(New feature as of Oniguruma 6.7.) 763 764 765### 28. ONIG_SYN_OP2_QMARK_BRACE_CALLOUT_CONTENTS (enable `(?{...})`) 766 767_Set in: Perl, Perl_NG, Oniguruma_ 768 769Enables support for Perl-style "callouts" — pattern substitutions that result from 770invoking a callback method. When `(?{foo})` is reached in a pattern, the callback 771function set in `onig_set_progress_callout()` will be invoked, and be able to perform 772custom computation during the pattern match (and during backtracking). 773 774Full documentation for this advanced feature can be found in the Oniguruma 775`docs/CALLOUT.md` file, with an example in `samples/callout.c`. 776 777(New feature as of Oniguruma 6.8.) 778 779 780### 29. ONIG_SYN_OP2_ASTERISK_CALLOUT_NAME (enable `(*name)`) 781 782_Set in: Perl, Perl_NG, Oniguruma_ 783 784Enables support for Perl-style "callouts" — pattern substitutions that result from 785invoking a callback method. When `(*foo)` is reached in a pattern, the callback 786function set in `onig_set_callout_of_name()` will be invoked, passing the given name 787`foo` to it, and it can perform custom computation during the pattern match (and 788during backtracking). 789 790Full documentation for this advanced feature can be found in the Oniguruma 791`docs/CALLOUT.md` file, with an example in `samples/callout.c`. 792 793(New feature as of Oniguruma 6.8.) 794 795 796### 30. ONIG_SYN_OP2_OPTION_ONIGURUMA (enable options `(?imxWSDPy)` and `(?-imxWDSP)`) 797 798_Set in: Oniguruma_ 799 800Enables support of regex options. (i,m,x,W,S,D,P,y) 801 802(New feature as of Oniguruma 6.9.2) 803 804 - `i` - Case-insensitivity 805 - `m` - Multi-line mode (`.` can match `\n`) 806 - `x` - Extended pattern (free-formatting: whitespace will ignored) 807 - `W` - ASCII only word. 808 - `D` - ASCII only digit. 809 - `S` - ASCII only space. 810 - `P` - ASCII only POSIX properties. (includes W,D,S) 811 812---------- 813 814 815## Syntax Flags (syn) 816 817 818This group contains rules to handle corner cases and constructs that are errors in 819some syntaxes but not in others. 820 821### 0. ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS (independent `?`, `*`, `+`, `{n,m}`) 822 823_Set in: PosixExtended, GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 824 825This flag specifies how to handle operators like `?` and `*` when they aren't 826directly attached to an operand, as in `^*` or `(*)`: Are they an error, are 827they discarded, or are they taken as literals? If this flag is clear, they 828are taken as literals; otherwise, the ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS flag 829determines if they are errors or if they are discarded. 830 831### 1. ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS (error or ignore independent operators) 832 833_Set in: PosixExtended, GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 834 835If ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS is set, this flag controls what happens when 836independent operators appear in a pattern: If this flag is set, then independent 837operators produce an error message; if this flag is clear, then independent 838operators are silently discarded. 839 840### 2. ONIG_SYN_ALLOW_UNMATCHED_CLOSE_SUBEXP (allow `...)...`) 841 842_Set in: PosixExtended_ 843 844This flag, if set, causes a `)` character without a preceding `(` to be treated as 845a literal `)`, equivalent to `\)`. If this flag is clear, then an unmatched `)` 846character will produce an error message. 847 848### 3. ONIG_SYN_ALLOW_INVALID_INTERVAL (allow `{???`) 849 850_Set in: GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 851 852This flag, if set, causes an invalid range, like `foo{bar}` or `foo{}`, to be 853silently discarded, as if `foo` had been written instead. If clear, an invalid 854range will produce an error message. 855 856### 4. ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV (allow `{,n}` to mean `{0,n}`) 857 858_Set in: Ruby, Oniguruma_ 859 860If this flag is set, then `r{,n}` will be treated as equivalent to writing 861`{0,n}`. If this flag is clear, then `r{,n}` will produce an error message. 862 863Note that regardless of whether this flag is set or clear, if 864ONIG_SYN_OP_BRACE_INTERVAL is enabled, then `r{n,}` will always be legal: This 865flag *only* controls the behavior of the opposite form, `r{,n}`. 866 867### 5. ONIG_SYN_STRICT_CHECK_BACKREF (error on invalid backrefs) 868 869_Set in: none_ 870 871If this flag is set, an invalid backref, like `\1` in a pattern with no captures, 872will produce an error. If this flag is clear, then an invalid backref will be 873equivalent to the empty string. 874 875No built-in syntax has this flag enabled. 876 877### 6. ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND (allow `(?<=a|bc)`) 878 879_Set in: Java, Ruby, Oniguruma_ 880 881If this flag is set, lookbehind patterns with alternate options may have differing 882lengths among those options. If this flag is clear, lookbehind patterns with options 883must have each option have identical length to the other options. 884 885Oniguruma can handle either form, but not all regex engines can, so for compatibility, 886Oniguruma allows you to cause regexes for other regex engines to fail if they might 887depend on this rule. 888 889### 7. ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP (prefer `\k<name>` over `\3`) 890 891_Set in: Perl_NG, Ruby, Oniguruma_ 892 893If this flag is set on the syntax *and* ONIG_OPTION_CAPTURE_GROUP is set when calling 894Oniguruma, then if a name is used on any capture, all captures must also use names: A 895single use of a named capture prohibits the use of numbered captures. 896 897### 8. ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME (allow `(?<x>)...(?<x>)`) 898 899_Set in: Perl_NG, Ruby, Oniguruma_ 900 901If this flag is set, multiple capture groups may use the same name. If this flag is 902clear, then reuse of a name will produce an error message. 903 904### 9. ONIG_SYN_FIXED_INTERVAL_IS_GREEDY_ONLY (`a{n}?` is equivalent to `(?:a{n})?`) 905 906_Set in: Ruby, Oniguruma_ 907 908If this flag is set, then intervals of a fixed size will ignore a lazy (non-greedy) 909`?` quantifier and treat it as an optional match (an ordinary `r?`), since "match as 910little as possible" is meaningless for a fixed-size interval. If this flag is clear, 911then `r{n}?` will mean the same as `r{n}`, and the useless `?` will be discarded. 912 913### 20. ONIG_SYN_NOT_NEWLINE_IN_NEGATIVE_CC (add `\n` to `[^...]`) 914 915_Set in: Grep_ 916 917If this flag is set, all newline characters (like `\n`) will be excluded from a negative 918character class automatically, as if the pattern had been written as `[^...\n]`. If this 919flag is clear, negative character classes do not automatically exclude newlines, and 920only exclude those characters and ranges written in them. 921 922### 21. ONIG_SYN_BACKSLASH_ESCAPE_IN_CC (allow `[...\w...]`) 923 924_Set in: GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 925 926If this flag is set, shorthands like `\w` are allowed to describe characters in character 927classes. If this flag is clear, shorthands like `\w` are treated as a redundantly-escaped 928literal `w`. 929 930### 22. ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC (silently discard `[z-a]`) 931 932_Set in: Emacs, Grep_ 933 934If this flag is set, then character ranges like `[z-a]` that are broken or contain no 935characters will be silently ignored. If this flag is clear, then broken or empty 936character ranges will produce an error message. 937 938### 23. ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC (treat `[0-9-a]` as `[0-9\-a]`) 939 940_Set in: PosixExtended, GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 941 942If this flag is set, then a trailing `-` after a character range will be taken as a 943literal `-`, as if it had been escaped as `\-`. If this flag is clear, then a trailing 944`-` after a character range will produce an error message. 945 946### 24. ONIG_SYN_WARN_CC_OP_NOT_ESCAPED (warn on `[[...]` and `[-x]`) 947 948_Set in: Ruby, Oniguruma_ 949 950If this flag is set, Oniguruma will be stricter about warning for bad forms in 951character classes: `[[...]` will produce a warning, but `[\[...]` will not; 952`[-x]` will produce a warning, but `[\-x]` will not; `[x&&-y]` will produce a warning, 953while `[x&&\-y]` will not; and so on. If this flag is clear, all of these warnings 954will be silently discarded. 955 956### 25. ONIG_SYN_WARN_REDUNDANT_NESTED_REPEAT (warn on `(?:a*)+`) 957 958_Set in: Ruby, Oniguruma_ 959 960If this flag is set, Oniguruma will warn about nested repeat operators those have no meaning, like `(?:a*)+`. 961If this flag is clear, Oniguruma will allow the nested repeat operators without warning about them. 962 963### 26. ONIG_SYN_ALLOW_INVALID_CODE_END_OF_RANGE_IN_CC (allow [a-\x{7fffffff}]) 964 965_Set in: Oniguruma_ 966 967If this flag is set, then invalid code points at the end of range in character class are allowed. 968 969### 31. ONIG_SYN_CONTEXT_INDEP_ANCHORS 970 971_Set in: PosixExtended, GnuRegex, Java, Perl, Perl_NG, Ruby, Oniguruma_ 972 973Not currently used, and does nothing. (But still set in several syntaxes for some 974reason.) 975 976---------- 977 978## Usage tables 979 980These tables show which of the built-in syntaxes use which flags and options, for easy comparison between them. 981 982### Group One Flags (op) 983 984| ID | Option | PosB | PosEx | Emacs | Grep | Gnu | Java | Perl | PeNG | Ruby | Onig | 985| ----- | --------------------------------------------- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | 986| 0 | `ONIG_SYN_OP_VARIABLE_META_CHARACTERS` | - | - | - | - | - | - | - | - | - | - | 987| 1 | `ONIG_SYN_OP_DOT_ANYCHAR` | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 988| 2 | `ONIG_SYN_OP_ASTERISK_ZERO_INF` | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 989| 3 | `ONIG_SYN_OP_ESC_ASTERISK_ZERO_INF` | - | - | - | - | - | - | - | - | - | - | 990| 4 | `ONIG_SYN_OP_PLUS_ONE_INF` | - | Yes | Yes | - | Yes | Yes | Yes | Yes | Yes | Yes | 991| 5 | `ONIG_SYN_OP_ESC_PLUS_ONE_INF` | - | - | - | Yes | - | - | - | - | - | - | 992| 6 | `ONIG_SYN_OP_QMARK_ZERO_ONE` | - | Yes | Yes | - | Yes | Yes | Yes | Yes | Yes | Yes | 993| 7 | `ONIG_SYN_OP_ESC_QMARK_ZERO_ONE` | - | - | - | Yes | - | - | - | - | - | - | 994| 8 | `ONIG_SYN_OP_BRACE_INTERVAL` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 995| 9 | `ONIG_SYN_OP_ESC_BRACE_INTERVAL` | Yes | - | Yes | Yes | - | - | - | - | - | - | 996| 10 | `ONIG_SYN_OP_VBAR_ALT` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 997| 11 | `ONIG_SYN_OP_ESC_VBAR_ALT` | - | - | Yes | Yes | - | - | - | - | - | - | 998| 12 | `ONIG_SYN_OP_LPAREN_SUBEXP` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 999| 13 | `ONIG_SYN_OP_ESC_LPAREN_SUBEXP` | Yes | - | Yes | Yes | - | - | - | - | - | - | 1000| 14 | `ONIG_SYN_OP_ESC_AZ_BUF_ANCHOR` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1001| 15 | `ONIG_SYN_OP_ESC_CAPITAL_G_BEGIN_ANCHOR` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1002| 16 | `ONIG_SYN_OP_DECIMAL_BACKREF` | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 1003| 17 | `ONIG_SYN_OP_BRACKET_CC` | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 1004| 18 | `ONIG_SYN_OP_ESC_W_WORD` | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 1005| 19 | `ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END` | - | - | - | Yes | Yes | - | - | - | - | - | 1006| 20 | `ONIG_SYN_OP_ESC_B_WORD_BOUND` | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 1007| 21 | `ONIG_SYN_OP_ESC_S_WHITE_SPACE` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1008| 22 | `ONIG_SYN_OP_ESC_D_DIGIT` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1009| 23 | `ONIG_SYN_OP_LINE_ANCHOR` | - | - | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 1010| 24 | `ONIG_SYN_OP_POSIX_BRACKET` | Yes | Yes | Yes | - | Yes | Yes | Yes | Yes | Yes | Yes | 1011| 25 | `ONIG_SYN_OP_QMARK_NON_GREEDY` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1012| 26 | `ONIG_SYN_OP_ESC_CONTROL_CHARS` | Yes | Yes | - | - | - | Yes | Yes | Yes | Yes | Yes | 1013| 27 | `ONIG_SYN_OP_ESC_C_CONTROL` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1014| 28 | `ONIG_SYN_OP_ESC_OCTAL3` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1015| 29 | `ONIG_SYN_OP_ESC_X_HEX2` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1016| 30 | `ONIG_SYN_OP_ESC_X_BRACE_HEX8` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1017| 31 | `ONIG_SYN_OP_ESC_O_BRACE_OCTAL` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1018 1019### Group Two Flags (op2) 1020 1021| ID | Option | PosB | PosEx | Emacs | Grep | Gnu | Java | Perl | PeNG | Ruby | Onig | 1022| ----- | --------------------------------------------- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | 1023| 0 | `ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE` | - | - | - | - | - | Yes | Yes | Yes | - | - | 1024| 1 | `ONIG_SYN_OP2_QMARK_GROUP_EFFECT` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1025| 2 | `ONIG_SYN_OP2_OPTION_PERL` | - | - | - | - | - | Yes | Yes | Yes | - | - | 1026| 3 | `ONIG_SYN_OP2_OPTION_RUBY` | - | - | - | - | - | - | - | - | Yes | - | 1027| 4 | `ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT` | - | - | - | - | - | - | - | - | Yes | Yes | 1028| 5 | `ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL` | - | - | - | - | - | Yes | - | - | - | - | 1029| 6 | `ONIG_SYN_OP2_CCLASS_SET_OP` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1030| 7 | `ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1031| 8 | `ONIG_SYN_OP2_ESC_K_NAMED_BACKREF` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1032| 9 | `ONIG_SYN_OP2_ESC_G_SUBEXP_CALL` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1033| 10 | `ONIG_SYN_OP2_ATMARK_CAPTURE_HISTORY` | - | - | - | - | - | - | - | - | - | - | 1034| 11 | `ONIG_SYN_OP2_ESC_CAPITAL_C_BAR_CONTROL` | - | - | - | - | - | - | - | - | Yes | Yes | 1035| 12 | `ONIG_SYN_OP2_ESC_CAPITAL_M_BAR_META` | - | - | - | - | - | - | - | - | Yes | Yes | 1036| 13 | `ONIG_SYN_OP2_ESC_V_VTAB` | - | - | - | - | - | Yes | - | - | Yes | Yes | 1037| 14 | `ONIG_SYN_OP2_ESC_U_HEX4` | - | - | - | - | - | Yes | - | - | Yes | Yes | 1038| 15 | `ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR` | - | - | Yes | - | - | - | - | - | - | - | 1039| 16 | `ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY` | - | - | - | - | - | Yes | Yes | Yes | Yes | Yes | 1040| 17 | `ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1041| 18 | `ONIG_SYN_OP2_CHAR_PROPERTY_PREFIX_IS` | - | - | - | - | - | - | - | - | - | - | 1042| 19 | `ONIG_SYN_OP2_ESC_H_XDIGIT` | - | - | - | - | - | - | - | - | Yes | Yes | 1043| 20 | `ONIG_SYN_OP2_INEFFECTIVE_ESCAPE` | - | - | - | - | - | - | - | - | - | - | 1044| 21 | `ONIG_SYN_OP2_QMARK_LPAREN_IF_ELSE` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1045| 22 | `ONIG_SYN_OP2_ESC_CAPITAL_K_KEEP` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1046| 23 | `ONIG_SYN_OP2_ESC_CAPITAL_R_GENERAL_NEWLINE` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1047| 24 | `ONIG_SYN_OP2_ESC_CAPITAL_N_O_SUPER_DOT` | - | - | - | - | - | - | Yes | Yes | - | Yes | 1048| 25 | `ONIG_SYN_OP2_QMARK_TILDE_ABSENT_GROUP` | - | - | - | - | - | - | - | - | Yes | Yes | 1049| 26 | `ONIG_SYN_OP2_ESC_X_Y_TEXT_SEGMENT` | - | - | - | - | - | - | Yes | Yes | Yes | Yes | 1050| 27 | `ONIG_SYN_OP2_QMARK_PERL_SUBEXP_CALL` | - | - | - | - | - | - | - | Yes | - | - | 1051| 28 | `ONIG_SYN_OP2_QMARK_BRACE_CALLOUT_CONTENTS` | - | - | - | - | - | - | Yes | Yes | Yes | - | 1052| 29 | `ONIG_SYN_OP2_ASTERISK_CALLOUT_NAME` | - | - | - | - | - | - | Yes | Yes | Yes | - | 1053| 30 | `ONIG_SYN_OP2_OPTION_ONIGURUMA` | - | - | - | - | - | - | - | - | - | Yes | 1054 1055### Syntax Flags (syn) 1056 1057| ID | Option | PosB | PosEx | Emacs | Grep | Gnu | Java | Perl | PeNG | Ruby | Onig | 1058| ----- | --------------------------------------------- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- | 1059| 0 | `ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1060| 1 | `ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1061| 2 | `ONIG_SYN_ALLOW_UNMATCHED_CLOSE_SUBEXP` | - | Yes | - | - | - | - | - | - | - | - | 1062| 3 | `ONIG_SYN_ALLOW_INVALID_INTERVAL` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1063| 4 | `ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV` | - | - | - | - | - | - | - | - | Yes | Yes | 1064| 5 | `ONIG_SYN_STRICT_CHECK_BACKREF` | - | - | - | - | - | - | - | - | - | - | 1065| 6 | `ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1066| 7 | `ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1067| 8 | `ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME` | - | - | - | - | - | - | - | Yes | Yes | Yes | 1068| 9 | `ONIG_SYN_FIXED_INTERVAL_IS_GREEDY_ONLY` | - | - | - | - | - | - | - | - | Yes | Yes | 1069| 20 | `ONIG_SYN_NOT_NEWLINE_IN_NEGATIVE_CC` | - | - | - | Yes | - | - | - | - | - | - | 1070| 21 | `ONIG_SYN_BACKSLASH_ESCAPE_IN_CC` | - | - | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1071| 22 | `ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC` | - | - | Yes | Yes | - | - | - | - | - | - | 1072| 23 | `ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1073| 24 | `ONIG_SYN_WARN_CC_OP_NOT_ESCAPED` | - | - | - | - | - | - | - | - | Yes | Yes | 1074| 25 | `ONIG_SYN_WARN_REDUNDANT_NESTED_REPEAT` | - | - | - | - | - | - | - | - | Yes | Yes | 1075| 26 | `ONIG_SYN_ALLOW_INVALID_CODE_END_OF_RANGE_IN_CC` | - | - | - | - | - | - | - | - | - | Yes | 1076| 31 | `ONIG_SYN_CONTEXT_INDEP_ANCHORS` | - | Yes | - | - | Yes | Yes | Yes | Yes | Yes | Yes | 1077
