| 1 | // Copyright (c) 2021 Lars Pontoppidan. All rights reserved. |
| 2 | // Use of this source code is governed by an MIT license |
| 3 | // that can be found in the LICENSE file. |
| 4 | module parser |
| 5 | |
| 6 | import toml.ast |
| 7 | import toml.checker |
| 8 | import toml.decoder |
| 9 | import toml.util |
| 10 | import toml.token |
| 11 | import toml.scanner |
| 12 | |
| 13 | pub const all_formatting = [token.Kind.whitespace, .tab, .cr, .nl] |
| 14 | pub const space_formatting = [token.Kind.whitespace, .tab] |
| 15 | pub const keys_and_space_formatting = [token.Kind.whitespace, .tab, .minus, .bare, .quoted, .boolean, |
| 16 | .number, .underscore] |
| 17 | |
| 18 | fn all_formatting_kinds() []token.Kind { |
| 19 | return [token.Kind.whitespace, .tab, .cr, .nl] |
| 20 | } |
| 21 | |
| 22 | fn space_formatting_kinds() []token.Kind { |
| 23 | return [token.Kind.whitespace, .tab] |
| 24 | } |
| 25 | |
| 26 | fn keys_and_space_formatting_kinds() []token.Kind { |
| 27 | return [token.Kind.whitespace, .tab, .minus, .bare, .quoted, .boolean, .number, .underscore] |
| 28 | } |
| 29 | |
| 30 | type DottedKey = []string |
| 31 | |
| 32 | // str returns the dotted key as a string. |
| 33 | pub fn (dk DottedKey) str() string { |
| 34 | return dk.join('.') |
| 35 | } |
| 36 | |
| 37 | // starts_with returns true if the dotted key starts with the same key entries as `target`. |
| 38 | fn (dk DottedKey) starts_with(target DottedKey) bool { |
| 39 | if dk.len >= target.len { |
| 40 | for i := 0; i < target.len; i++ { |
| 41 | if dk[i] != target[i] { |
| 42 | return false |
| 43 | } |
| 44 | } |
| 45 | return true |
| 46 | } |
| 47 | return false |
| 48 | } |
| 49 | |
| 50 | // has returns true if the array contains `target`. |
| 51 | fn (a []DottedKey) has(target DottedKey) bool { |
| 52 | for dk in a { |
| 53 | if dk == target { |
| 54 | return true |
| 55 | } |
| 56 | } |
| 57 | return false |
| 58 | } |
| 59 | |
| 60 | // Parser contains the necessary fields for keeping the state of the parsing process. |
| 61 | pub struct Parser { |
| 62 | pub: |
| 63 | config Config |
| 64 | mut: |
| 65 | scanner &scanner.Scanner = unsafe { nil } |
| 66 | prev_tok token.Token |
| 67 | tok token.Token |
| 68 | peek_tok token.Token |
| 69 | tokens []token.Token // To be able to peek more than one token ahead. |
| 70 | skip_next bool |
| 71 | // The root map (map is called table in TOML world) |
| 72 | root_map map[string]ast.Value |
| 73 | root_map_key DottedKey |
| 74 | value_is_immutable bool |
| 75 | immutable []DottedKey |
| 76 | explicit_declared []DottedKey |
| 77 | explicit_declared_array_of_tables []DottedKey |
| 78 | implicit_declared []DottedKey |
| 79 | // Array of Tables state |
| 80 | last_aot DottedKey |
| 81 | last_aot_index int |
| 82 | // Root of the tree |
| 83 | ast_root &ast.Root = &ast.Root{} |
| 84 | } |
| 85 | |
| 86 | // Config is used to configure a Parser instance. |
| 87 | // `run_checks` is used to en- or disable running of the strict `checker.Checker` type checks. |
| 88 | // `decode_values` is used to en- or disable decoding of values with the `decoder.Decoder`. |
| 89 | pub struct Config { |
| 90 | pub: |
| 91 | scanner &scanner.Scanner = unsafe { nil } |
| 92 | run_checks bool = true |
| 93 | decode_values bool = true |
| 94 | } |
| 95 | |
| 96 | // new_parser returns a new, stack allocated, `Parser`. |
| 97 | pub fn new_parser(config Config) Parser { |
| 98 | return Parser{ |
| 99 | config: config |
| 100 | scanner: config.scanner |
| 101 | } |
| 102 | } |
| 103 | |
| 104 | // init initializes the parser. |
| 105 | pub fn (mut p Parser) init() ! { |
| 106 | p.root_map = map[string]ast.Value{} |
| 107 | p.tokens << p.scanner.scan()! |
| 108 | p.next()! |
| 109 | } |
| 110 | |
| 111 | // run_checker validates the parsed `ast.Value` nodes in the |
| 112 | // the generated AST. |
| 113 | fn (mut p Parser) run_checker() ! { |
| 114 | if p.config.run_checks { |
| 115 | chckr := checker.Checker{ |
| 116 | scanner: p.scanner |
| 117 | } |
| 118 | chckr.check(p.root_map)! |
| 119 | for comment in p.ast_root.comments { |
| 120 | chckr.check_comment(comment)! |
| 121 | } |
| 122 | } |
| 123 | } |
| 124 | |
| 125 | // run_decoder decodes values in the parsed `ast.Value` nodes in the |
| 126 | // the generated AST. |
| 127 | fn (mut p Parser) run_decoder() ! { |
| 128 | if p.config.decode_values { |
| 129 | dcoder := decoder.Decoder{ |
| 130 | scanner: p.scanner |
| 131 | } |
| 132 | dcoder.decode(mut p.root_map)! |
| 133 | } |
| 134 | } |
| 135 | |
| 136 | // parse starts parsing the input and returns the root |
| 137 | // of the generated AST. |
| 138 | pub fn (mut p Parser) parse() !&ast.Root { |
| 139 | p.init()! |
| 140 | p.root_table()! |
| 141 | p.run_checker()! |
| 142 | p.run_decoder()! |
| 143 | p.ast_root.table = p.root_map |
| 144 | return p.ast_root |
| 145 | } |
| 146 | |
| 147 | // next forwards the parser to the next token. |
| 148 | fn (mut p Parser) next() ! { |
| 149 | p.prev_tok = p.tok |
| 150 | p.tok = p.peek_tok |
| 151 | if p.tokens.len > 0 { |
| 152 | p.peek_tok = p.tokens.first() |
| 153 | p.tokens.delete(0) |
| 154 | p.peek(1)! |
| 155 | } else { |
| 156 | p.peek(1)! |
| 157 | p.peek_tok = p.tokens.first() |
| 158 | p.tokens.delete(0) |
| 159 | } |
| 160 | } |
| 161 | |
| 162 | // peek peeks forward `n` tokens. |
| 163 | // peek returns `.unknown` if it can not peek ahead long enough. |
| 164 | fn (mut p Parser) peek(n int) !token.Token { |
| 165 | if n < 0 { |
| 166 | return error(@MOD + '.' + @STRUCT + '.' + @FN + ' peeking backwards is not supported.') |
| 167 | } |
| 168 | if n == 0 { |
| 169 | return p.peek_tok |
| 170 | } else { |
| 171 | // n >= 1 |
| 172 | if n <= p.tokens.len { |
| 173 | return p.tokens[n - 1] |
| 174 | } else { |
| 175 | mut token_ := token.Token{} |
| 176 | mut count := n - p.tokens.len |
| 177 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'buffering ${count} tokens...') |
| 178 | for token_.kind != .eof && count != 0 { |
| 179 | token_ = p.scanner.scan()! |
| 180 | p.tokens << token_ |
| 181 | count-- |
| 182 | } |
| 183 | return token_ |
| 184 | } |
| 185 | } |
| 186 | } |
| 187 | |
| 188 | // check forwards the parser to the next token if the current |
| 189 | // token's `Kind` is equal that of `check_token`. |
| 190 | fn (mut p Parser) check(check_token token.Kind) ! { |
| 191 | if p.tok.kind == check_token { |
| 192 | p.next()! |
| 193 | } else { |
| 194 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 195 | ' expected token "${check_token}" but found "${p.tok.kind}" in this (excerpt): "...${p.excerpt()}..."') |
| 196 | } |
| 197 | } |
| 198 | |
| 199 | // peek_for_correct_line_ending_or_fail peeks past any formatting tokens |
| 200 | // and return an error if the next token is not one of [.cr, .nl, .hash, .eof]. |
| 201 | fn (mut p Parser) peek_for_correct_line_ending_or_fail() ! { |
| 202 | // Disallow anything else than [.cr, .nl, .hash, .eof] after any space formatting. |
| 203 | peek_tok, _ := p.peek_over(1, space_formatting_kinds())! |
| 204 | if peek_tok.kind !in [.cr, .nl, .hash, .eof] { |
| 205 | p.next()! // Forward to the peek_tok |
| 206 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 207 | ' unexpected EOL "${p.tok.kind}" "${p.tok.lit}" expected one of [.cr, .nl, .hash, .eof] at this (excerpt): "...${p.excerpt()}..."') |
| 208 | } |
| 209 | } |
| 210 | |
| 211 | // check_one_of forwards the parser to the next token if the current |
| 212 | // token's `Kind` can be found in `tokens`. Otherwise it returns an error. |
| 213 | fn (mut p Parser) check_one_of(tokens []token.Kind) ! { |
| 214 | if p.tok.kind in tokens { |
| 215 | p.next()! |
| 216 | } else { |
| 217 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 218 | ' expected one of ${tokens} but found "${p.tok.kind}" in this (excerpt): "...${p.excerpt()}..."') |
| 219 | } |
| 220 | } |
| 221 | |
| 222 | // ignore_while forwards the parser to the next token as long as the current |
| 223 | // token's `Kind` can be found in `tokens`. This is helpful for ignoring |
| 224 | // a stream of formatting tokens. |
| 225 | fn (mut p Parser) ignore_while(tokens []token.Kind) { |
| 226 | if p.tok.kind in tokens { |
| 227 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'ignoring "${p.tok.kind}" ...') |
| 228 | p.next() or { return } |
| 229 | p.ignore_while(tokens) |
| 230 | } |
| 231 | } |
| 232 | |
| 233 | // ignore_while_peek forwards the parser to the next token as long as `peek_tok` |
| 234 | // token's `Kind` can be found in `tokens`. This is helpful for ignoring |
| 235 | // a stream of formatting tokens. |
| 236 | // In contrast to `ignore_while`, `ignore_while_peek` compares on `peek_tok` this is |
| 237 | // sometimes necessary since not all parser calls forward using the `next()` call. |
| 238 | fn (mut p Parser) ignore_while_peek(tokens []token.Kind) { |
| 239 | for p.peek_tok.kind in tokens { |
| 240 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'eating "${p.tok.kind}" ...') |
| 241 | p.next() or { return } |
| 242 | } |
| 243 | } |
| 244 | |
| 245 | // peek_over peeks ahead from token starting at `i` skipping over |
| 246 | // any `token.Kind`s found in `tokens`. `peek_over` returns the next token *not* |
| 247 | // found in `tokens`. |
| 248 | fn (mut p Parser) peek_over(i int, tokens []token.Kind) !(token.Token, int) { |
| 249 | mut peek_tok := p.peek_tok |
| 250 | |
| 251 | // Peek ahead as far as we can from token at `i` while the peeked |
| 252 | // token is found in `tokens`. |
| 253 | mut peek_i := i |
| 254 | for peek_tok.kind in tokens { |
| 255 | peek_tok = p.peek(peek_i)! |
| 256 | peek_i++ |
| 257 | } |
| 258 | return peek_tok, peek_i |
| 259 | } |
| 260 | |
| 261 | // is_at returns true if the token kind is equal to `expected_token`. |
| 262 | fn (mut p Parser) is_at(expected_token token.Kind) bool { |
| 263 | return p.tok.kind == expected_token |
| 264 | } |
| 265 | |
| 266 | // expect will error if the token kind is not equal to `expected_token`. |
| 267 | fn (mut p Parser) expect(expected_token token.Kind) ! { |
| 268 | if p.tok.kind == expected_token { |
| 269 | return |
| 270 | } else { |
| 271 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 272 | ' expected token "${expected_token}" but found "${p.tok.kind}" in this text "...${p.excerpt()}..."') |
| 273 | } |
| 274 | } |
| 275 | |
| 276 | // build_abs_dotted_key returns the absolute dotted key path. |
| 277 | fn (p &Parser) build_abs_dotted_key(key DottedKey) DottedKey { |
| 278 | if p.root_map_key.len > 0 { |
| 279 | mut abs_dotted_key := DottedKey([]string{}) |
| 280 | abs_dotted_key << p.root_map_key |
| 281 | abs_dotted_key << key |
| 282 | return abs_dotted_key |
| 283 | } |
| 284 | return key |
| 285 | } |
| 286 | |
| 287 | // todo_msvc_astring2dkey worksaround a MSVC compile error. |
| 288 | // TODO: remove. |
| 289 | fn todo_msvc_astring2dkey(s []string) DottedKey { |
| 290 | return s |
| 291 | } |
| 292 | |
| 293 | // check_immutable returns an error if `key` has been declared as immutable. |
| 294 | fn (p &Parser) check_immutable(key DottedKey) ! { |
| 295 | if p.immutable.len > 0 && p.immutable.has(key) { |
| 296 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 297 | ' key `${key.str()}` is immutable. Unexpected mutation at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 298 | } |
| 299 | } |
| 300 | |
| 301 | // check_explicitly_declared returns an error if `key` has been explicitly declared. |
| 302 | fn (p &Parser) check_explicitly_declared(key DottedKey) ! { |
| 303 | if p.explicit_declared.len > 0 && p.explicit_declared.has(key) { |
| 304 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 305 | ' key `${key.str()}` is already explicitly declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 306 | } |
| 307 | } |
| 308 | |
| 309 | // check_explicitly_declared_array_of_tables returns an error if `key` has been |
| 310 | // explicitly declared as an array of tables. |
| 311 | fn (p &Parser) check_explicitly_declared_array_of_tables(key DottedKey) ! { |
| 312 | if p.explicit_declared_array_of_tables.len > 0 && p.explicit_declared_array_of_tables.has(key) { |
| 313 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 314 | ' key `${key.str()}` is already an explicitly declared array of tables. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 315 | } |
| 316 | } |
| 317 | |
| 318 | // check_implicitly_declared returns an error if `key` has been implicitly declared. |
| 319 | fn (p &Parser) check_implicitly_declared(key DottedKey) ! { |
| 320 | if p.implicit_declared.len > 0 && p.implicit_declared.has(key) { |
| 321 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 322 | ' key `${key.str()}` is already implicitly declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 323 | } |
| 324 | } |
| 325 | |
| 326 | // find_table returns a reference to a map if found in the *root* table given a "dotted" key (`a.b.c`). |
| 327 | // If some segments of the key does not exist in the root table find_table will |
| 328 | // allocate a new map for each segment. This behavior is needed because you can |
| 329 | // reference maps by multiple keys "dotted" (separated by "." periods) in TOML documents. |
| 330 | // See also `find_in_table`. |
| 331 | pub fn (mut p Parser) find_table() !&map[string]ast.Value { |
| 332 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 333 | 'locating "${p.root_map_key}" in map ${ptr_str(p.root_map)}') |
| 334 | mut t := unsafe { &p.root_map } |
| 335 | if p.root_map_key.len == 0 { |
| 336 | return t |
| 337 | } |
| 338 | |
| 339 | return p.find_in_table(mut t, p.root_map_key) |
| 340 | } |
| 341 | |
| 342 | // allocate_table allocates all tables in "dotted" `key` (`a.b.c`) in the *root* table. |
| 343 | pub fn (mut p Parser) allocate_table(key DottedKey) ! { |
| 344 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 345 | 'allocating "${key}" in map ${ptr_str(p.root_map)}') |
| 346 | mut t := unsafe { &p.root_map } |
| 347 | if key.len == 0 { |
| 348 | return |
| 349 | } |
| 350 | p.allocate_in_table(mut t, key)! |
| 351 | } |
| 352 | |
| 353 | // sub_table_key returns the logic parts of a dotted key (`a.b.c`) for |
| 354 | // use with the `find_sub_table` method. |
| 355 | pub fn (mut p Parser) sub_table_key(key DottedKey) (DottedKey, DottedKey) { |
| 356 | last := [key.last()] |
| 357 | first := key[..key.len - 1] |
| 358 | return first, last |
| 359 | } |
| 360 | |
| 361 | // find_sub_table returns a reference to a map if found in the *root* table given a "dotted" key (`a.b.c`). |
| 362 | // If some segments of the key does not exist in the input map find_sub_table will |
| 363 | // allocate a new map for the segment. This behavior is needed because you can |
| 364 | // reference maps by multiple keys "dotted" (separated by "." periods) in TOML documents. |
| 365 | // See also `find_in_table`. |
| 366 | pub fn (mut p Parser) find_sub_table(key DottedKey) !&map[string]ast.Value { |
| 367 | mut ky := DottedKey([]string{}) |
| 368 | ky << p.root_map_key |
| 369 | ky << key |
| 370 | if p.root_map_key.len == 0 { |
| 371 | ky = unsafe { key } |
| 372 | } |
| 373 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 374 | 'locating "${ky}" in map ${ptr_str(p.root_map)}') |
| 375 | mut t := unsafe { &p.root_map } |
| 376 | if ky.len == 0 { |
| 377 | return t |
| 378 | } |
| 379 | |
| 380 | return p.find_in_table(mut t, ky) |
| 381 | } |
| 382 | |
| 383 | // find_in_table returns a reference to a map if found in `table` given a "dotted" key (`a.b.c`). |
| 384 | // If some segments of the key does not exist in the input map find_in_table will |
| 385 | // allocate a new map for the segment. This behavior is needed because you can |
| 386 | // reference maps by multiple keys "dotted" (separated by "." periods) in TOML documents. |
| 387 | pub fn (mut p Parser) find_in_table(mut table map[string]ast.Value, key DottedKey) !&map[string]ast.Value { |
| 388 | // NOTE This code is the result of much trial and error. |
| 389 | // I'm still not quite sure *exactly* why it works. All I can leave here is a hope |
| 390 | // that this kind of minefield someday will be easier in V :) |
| 391 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'locating "${key}" in map ${ptr_str(table)}') |
| 392 | mut t := unsafe { &table } |
| 393 | unsafe { |
| 394 | for k in key { |
| 395 | if val := t[k] { |
| 396 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'found key "${k}" in ${t.keys()}') |
| 397 | if val is map[string]ast.Value { |
| 398 | t = &val |
| 399 | } else { |
| 400 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 401 | ' "${k}" in "${key}" is not a map but `${val.type_name()}`') |
| 402 | } |
| 403 | } else { |
| 404 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 405 | 'no key "${k}" in "${key}" found, allocating new map at key "${k}" in map ${ptr_str(t)}"') |
| 406 | t[k] = map[string]ast.Value{} |
| 407 | t = &(t[k] as map[string]ast.Value) |
| 408 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'allocated new map ${ptr_str(t)}"') |
| 409 | } |
| 410 | } |
| 411 | } |
| 412 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'returning map ${ptr_str(t)}"') |
| 413 | return t |
| 414 | } |
| 415 | |
| 416 | // is_all_tables returns `true` if *all* entries in `dotted_key` (`a.b.c`) are tables (`map[string]ast.Value`), `false` otherwise. |
| 417 | fn is_all_tables(table map[string]ast.Value, dotted_key DottedKey) bool { |
| 418 | if dotted_key.len == 0 { |
| 419 | return false |
| 420 | } |
| 421 | unsafe { |
| 422 | mut t := &table |
| 423 | for key in dotted_key { |
| 424 | if val := t[key] { |
| 425 | if val is map[string]ast.Value { |
| 426 | t = &val |
| 427 | } else { |
| 428 | return false |
| 429 | } |
| 430 | } else { |
| 431 | return false |
| 432 | } |
| 433 | } |
| 434 | } |
| 435 | return true |
| 436 | } |
| 437 | |
| 438 | // find_array_of_tables returns an array if found in the root table based on the parser's |
| 439 | // last encountered "Array Of Tables" key. |
| 440 | // If the state key does not exist find_array_in_table will return an error. |
| 441 | pub fn (mut p Parser) find_array_of_tables() ![]ast.Value { |
| 442 | mut t := unsafe { &p.root_map } |
| 443 | mut key := p.last_aot |
| 444 | if key.len > 1 { |
| 445 | key = DottedKey([key[0]]) |
| 446 | } |
| 447 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'locating "${key}" in map ${ptr_str(t)}') |
| 448 | unsafe { |
| 449 | if val := t[key.str()] { |
| 450 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'found key "${key}" in ${t.keys()}') |
| 451 | if val is []ast.Value { |
| 452 | return val |
| 453 | } |
| 454 | } |
| 455 | } |
| 456 | return error(@MOD + '.' + @STRUCT + '.' + @FN + 'no key `${key}` found in map ${ptr_str(t)}"') |
| 457 | } |
| 458 | |
| 459 | // allocate_in_table allocates all tables in "dotted" `key` (`a.b.c`) in `table`. |
| 460 | pub fn (mut p Parser) allocate_in_table(mut table map[string]ast.Value, key DottedKey) ! { |
| 461 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'allocating "${key}" in map ${ptr_str(table)}') |
| 462 | mut t := unsafe { &table } |
| 463 | unsafe { |
| 464 | for k in key { |
| 465 | if val := t[k] { |
| 466 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'found key "${k}" in ${t.keys()}') |
| 467 | if val is map[string]ast.Value { |
| 468 | t = &val |
| 469 | } else { |
| 470 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 471 | ' "${k}" in "${key}" is not a map (${val.type_name()})') |
| 472 | } |
| 473 | } else { |
| 474 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 475 | 'no key "${k}" in "${key}" found, allocating new map at key "${k}" in map ${ptr_str(t)}"') |
| 476 | t[k] = map[string]ast.Value{} |
| 477 | t = &(t[k] as map[string]ast.Value) |
| 478 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'allocated new map ${ptr_str(t)}"') |
| 479 | } |
| 480 | } |
| 481 | } |
| 482 | } |
| 483 | |
| 484 | // dotted_key returns a string of the next tokens parsed as |
| 485 | // sub/nested/path keys (e.g. `a.b.c`). In TOML, this form of key is referred to as a "dotted" key. |
| 486 | pub fn (mut p Parser) dotted_key() !DottedKey { |
| 487 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing dotted key...') |
| 488 | mut dotted_key := DottedKey([]string{}) |
| 489 | key := p.key()! |
| 490 | p.ignore_while_peek(space_formatting_kinds()) |
| 491 | dotted_key << key.str() |
| 492 | for p.peek_tok.kind == .period { |
| 493 | p.next()! // . |
| 494 | p.check(.period)! |
| 495 | p.ignore_while(space_formatting_kinds()) |
| 496 | next_key := p.key()! |
| 497 | dotted_key << next_key.text |
| 498 | p.ignore_while_peek(space_formatting_kinds()) |
| 499 | } |
| 500 | p.next()! |
| 501 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 502 | 'parsed dotted key `${dotted_key}` now at "${p.tok.kind}" "${p.tok.lit}"') |
| 503 | return dotted_key |
| 504 | } |
| 505 | |
| 506 | // root_table parses next tokens into the root map of `ast.Value`s. |
| 507 | // The V `map` type is corresponding to a "table" in TOML. |
| 508 | pub fn (mut p Parser) root_table() ! { |
| 509 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing root table...') |
| 510 | |
| 511 | for p.tok.kind != .eof { |
| 512 | if !p.skip_next { |
| 513 | p.next()! |
| 514 | } else { |
| 515 | p.skip_next = false |
| 516 | } |
| 517 | |
| 518 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 519 | 'parsing token "${p.tok.kind}" "${p.tok.lit}"') |
| 520 | match p.tok.kind { |
| 521 | .hash { |
| 522 | c := p.comment() |
| 523 | p.ast_root.comments << c |
| 524 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'skipping comment "${c.text}"') |
| 525 | } |
| 526 | .whitespace, .tab, .nl, .cr { |
| 527 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 528 | 'skipping formatting "${p.tok.kind}" "${p.tok.lit}"') |
| 529 | continue |
| 530 | } |
| 531 | .bare, .boolean, .quoted, .number, .minus, .underscore { |
| 532 | // Peek forward as far as we can skipping over space formatting tokens. |
| 533 | peek_tok, _ := p.peek_over(1, keys_and_space_formatting_kinds())! |
| 534 | |
| 535 | if peek_tok.kind == .period { |
| 536 | dotted_key, val := p.dotted_key_value()! |
| 537 | |
| 538 | sub_table, key := p.sub_table_key(dotted_key) |
| 539 | |
| 540 | if is_all_tables(p.root_map, dotted_key) { |
| 541 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 542 | ' key `${dotted_key.str()}` is already declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 543 | } |
| 544 | |
| 545 | // NOTE these are *relatively* costly checks. In general - and by specification, |
| 546 | // TOML documents are expected to be "small" so this shouldn't be a problem. Famous last words. |
| 547 | for explicit_key in p.explicit_declared { |
| 548 | // Check for key re-defining: |
| 549 | // https://github.com/iarna/toml-spec-tests/blob/1880b1a/errors/inline-table-imutable-1.toml |
| 550 | |
| 551 | if p.build_abs_dotted_key(sub_table) == explicit_key { |
| 552 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 553 | ' key `${sub_table}` has already been explicitly declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 554 | } |
| 555 | if explicit_key.len == 1 || explicit_key == p.root_map_key { |
| 556 | continue |
| 557 | } |
| 558 | // Check for "table injection": |
| 559 | // https://github.com/toml-lang/toml-test/blob/576db85/tests/invalid/table/injection-1.toml |
| 560 | // https://github.com/toml-lang/toml-test/blob/576db85/tests/invalid/table/injection-2.toml |
| 561 | if p.build_abs_dotted_key(sub_table).starts_with(explicit_key) { |
| 562 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 563 | ' key `${dotted_key}` has already been explicitly declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 564 | } |
| 565 | } |
| 566 | |
| 567 | // Register implicit declaration |
| 568 | mut dotted_key_copy := dotted_key.clone() |
| 569 | dotted_key_copy.pop() |
| 570 | implicit_keys := todo_msvc_astring2dkey(dotted_key_copy) |
| 571 | mut abs_dotted_key := p.build_abs_dotted_key(implicit_keys) |
| 572 | if !p.implicit_declared.has(abs_dotted_key) { |
| 573 | p.implicit_declared << abs_dotted_key |
| 574 | } |
| 575 | |
| 576 | t := p.find_sub_table(sub_table)! |
| 577 | unsafe { |
| 578 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 579 | 'setting "${key}" = ${val} in table ${ptr_str(t)}') |
| 580 | t[key.str()] = val |
| 581 | } |
| 582 | } else { |
| 583 | p.ignore_while(space_formatting_kinds()) |
| 584 | key, val := p.key_value()! |
| 585 | |
| 586 | t := p.find_table()! |
| 587 | unsafe { |
| 588 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 589 | 'setting "${key.str()}" = ${val} in table ${ptr_str(t)}') |
| 590 | key_str := key.str() |
| 591 | if _ := t[key_str] { |
| 592 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 593 | ' key "${key}" is already initialized with a value. At "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 594 | } |
| 595 | t[key_str] = val |
| 596 | } |
| 597 | } |
| 598 | p.peek_for_correct_line_ending_or_fail()! |
| 599 | } |
| 600 | .lsbr { |
| 601 | p.check(.lsbr)! // '[' bracket |
| 602 | mut peek_tok := p.peek_tok |
| 603 | |
| 604 | // Disallow `[ [table]]` |
| 605 | if p.tok.kind in space_formatting_kinds() { |
| 606 | peek_tok, _ = p.peek_over(1, space_formatting_kinds())! |
| 607 | if peek_tok.kind == .lsbr { |
| 608 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 609 | ' unexpected "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 610 | } |
| 611 | } |
| 612 | |
| 613 | // Allow `[ d.e.f]` |
| 614 | p.ignore_while(space_formatting_kinds()) |
| 615 | |
| 616 | // Peek forward as far as we can skipping over space formatting tokens. |
| 617 | peek_tok, _ = p.peek_over(1, keys_and_space_formatting_kinds())! |
| 618 | |
| 619 | if p.tok.kind == .lsbr { |
| 620 | // Parse `[[table]]` |
| 621 | unsafe { |
| 622 | p.array_of_tables(mut &p.root_map)! |
| 623 | } |
| 624 | p.skip_next = true // skip calling p.next() in coming iteration |
| 625 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 626 | 'leaving double bracket at "${p.tok.kind}" "${p.tok.lit}". NEXT is "${p.peek_tok.kind} "${p.peek_tok.lit}"') |
| 627 | } else if peek_tok.kind == .period { |
| 628 | // Parse `[d.e.f]` |
| 629 | dotted_key := p.dotted_key()! |
| 630 | |
| 631 | // So apparently TOML is a *very* key context sensitive language... |
| 632 | // [[table]] <- parsed previously |
| 633 | // ... |
| 634 | // [table.key] <- parser is here |
| 635 | // |
| 636 | // `table.key` now shape shifts into being a *double array of tables* key... |
| 637 | // ... but with a different set of rules - making it hard to reuse the code we already have for that ... |
| 638 | // See `testdata/array_of_tables_edge_case_<N>_test.toml` for the type of constructs parsed. |
| 639 | if p.last_aot.len == 1 && dotted_key.len > 1 |
| 640 | && dotted_key[0] == p.last_aot.str() { |
| 641 | // Disallow re-declaring the key |
| 642 | p.check_explicitly_declared_array_of_tables(dotted_key)! |
| 643 | p.check(.rsbr)! |
| 644 | p.ignore_while(space_formatting_kinds()) |
| 645 | arr := p.find_array_of_tables()! |
| 646 | if val := arr[p.last_aot_index] { |
| 647 | if val is map[string]ast.Value { |
| 648 | mut m := map[string]ast.Value{} |
| 649 | p.table_contents(mut m)! |
| 650 | unsafe { |
| 651 | mut mut_val := &val |
| 652 | if dotted_key.len == 2 { |
| 653 | // [table.key] |
| 654 | mut_val[dotted_key[1].str()] = m |
| 655 | } else { |
| 656 | // [table.key.key.etc] |
| 657 | mut dotted_key_copy := dotted_key.clone() |
| 658 | dotted_key_copy.delete(0) |
| 659 | new_key := todo_msvc_astring2dkey(dotted_key_copy) |
| 660 | sub_table, key := p.sub_table_key(new_key) |
| 661 | t := p.find_in_table(mut mut_val, sub_table)! |
| 662 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 663 | 'setting "${key}" = ${val} in table ${ptr_str(t)}') |
| 664 | t[new_key.last().str()] = m |
| 665 | } |
| 666 | } |
| 667 | } else { |
| 668 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 669 | ' "${p.last_aot_index}" in array is not a map but `${typeof(val).name}`') |
| 670 | } |
| 671 | } |
| 672 | continue |
| 673 | } |
| 674 | |
| 675 | // Disallow mutation of immutable values (inline tables) |
| 676 | if dotted_key.len > 1 { |
| 677 | for part in dotted_key { |
| 678 | dotted_part := DottedKey([part]) |
| 679 | if p.explicit_declared.has(dotted_part) { |
| 680 | p.check_immutable(dotted_part)! |
| 681 | } |
| 682 | } |
| 683 | } |
| 684 | |
| 685 | // Disallow re-defining |
| 686 | // This check also covers *implicit* table allocations from "dotted" keys, so no need for e.g: `p.check_implicitly_declared(dotted_key)!` |
| 687 | if is_all_tables(p.root_map, dotted_key) { |
| 688 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 689 | ' key `${dotted_key.str()}` is already declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 690 | } |
| 691 | p.explicit_declared << dotted_key |
| 692 | |
| 693 | p.ignore_while(space_formatting_kinds()) |
| 694 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 695 | 'setting root map key to `${dotted_key}` at "${p.tok.kind}" "${p.tok.lit}"') |
| 696 | p.root_map_key = dotted_key |
| 697 | p.allocate_table(p.root_map_key)! |
| 698 | p.expect(.rsbr)! |
| 699 | p.peek_for_correct_line_ending_or_fail()! |
| 700 | } else { |
| 701 | // Parse `[key]` |
| 702 | key := p.key()! |
| 703 | dotted_key := DottedKey([key.str()]) |
| 704 | |
| 705 | p.check_implicitly_declared(dotted_key) or { |
| 706 | p.check_explicitly_declared(dotted_key) or { |
| 707 | if p.root_map[key.str()] or { ast.Bool{} } is map[string]ast.Value { |
| 708 | // NOTE: Here we "undo" the implicit-explicit special case declaration for: |
| 709 | // https://github.com/toml-lang/toml-test/blob/576db852/tests/invalid/table/array-implicit.toml |
| 710 | // ... to make the following test pass: |
| 711 | // https://github.com/toml-lang/toml-test/blob/229ce2e/tests/valid/table/array-implicit-and-explicit-after.toml |
| 712 | p.undo_special_case_01(dotted_key) |
| 713 | } |
| 714 | } |
| 715 | } |
| 716 | // Disallow re-declaring the key |
| 717 | p.check_explicitly_declared(dotted_key)! |
| 718 | p.explicit_declared << dotted_key |
| 719 | |
| 720 | // Allow [ key ] |
| 721 | p.ignore_while(space_formatting_kinds()) |
| 722 | |
| 723 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 724 | 'setting root map key to `${dotted_key}` at "${p.tok.kind}" "${p.tok.lit}"') |
| 725 | p.root_map_key = dotted_key |
| 726 | p.allocate_table(p.root_map_key)! |
| 727 | p.next()! |
| 728 | p.ignore_while(space_formatting_kinds()) |
| 729 | p.expect(.rsbr)! |
| 730 | p.peek_for_correct_line_ending_or_fail()! |
| 731 | } |
| 732 | } |
| 733 | .eof { |
| 734 | return |
| 735 | } |
| 736 | else { |
| 737 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 738 | ' could not parse "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 739 | } |
| 740 | } |
| 741 | } |
| 742 | } |
| 743 | |
| 744 | // excerpt returns a string of the characters surrounding `Parser.tok.pos` |
| 745 | fn (p &Parser) excerpt() string { |
| 746 | return p.scanner.excerpt(p.tok.pos, 10) |
| 747 | } |
| 748 | |
| 749 | // table_contents parses next tokens into a map of `ast.Value`s. |
| 750 | // The V `map` type is corresponding to a "table" in TOML. |
| 751 | pub fn (mut p Parser) table_contents(mut tbl map[string]ast.Value) ! { |
| 752 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing table contents...') |
| 753 | |
| 754 | for p.tok.kind != .eof { |
| 755 | if p.peek_tok.kind == .lsbr { |
| 756 | return |
| 757 | } |
| 758 | if !p.skip_next { |
| 759 | p.next()! |
| 760 | } else { |
| 761 | p.skip_next = false |
| 762 | } |
| 763 | |
| 764 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 765 | 'parsing token "${p.tok.kind}" "${p.tok.lit}"') |
| 766 | match p.tok.kind { |
| 767 | .hash { |
| 768 | c := p.comment() |
| 769 | p.ast_root.comments << c |
| 770 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'skipping comment "${c.text}"') |
| 771 | } |
| 772 | .whitespace, .tab, .nl, .cr { |
| 773 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 774 | 'skipping formatting "${p.tok.kind}" "${p.tok.lit}"') |
| 775 | continue |
| 776 | } |
| 777 | .bare, .quoted, .number, .minus, .underscore { |
| 778 | // Peek forward as far as we can skipping over space formatting tokens. |
| 779 | peek_tok, _ := p.peek_over(1, keys_and_space_formatting_kinds())! |
| 780 | |
| 781 | if peek_tok.kind == .period { |
| 782 | dotted_key, val := p.dotted_key_value()! |
| 783 | |
| 784 | sub_table, key := p.sub_table_key(dotted_key) |
| 785 | |
| 786 | t := p.find_in_table(mut tbl, sub_table)! |
| 787 | unsafe { |
| 788 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 789 | 'setting "${key}" = ${val} in table ${ptr_str(t)}') |
| 790 | t[key.str()] = val |
| 791 | } |
| 792 | } else { |
| 793 | p.ignore_while(space_formatting_kinds()) |
| 794 | key, val := p.key_value()! |
| 795 | |
| 796 | unsafe { |
| 797 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 798 | 'setting "${key.str()}" = ${val} in table ${ptr_str(tbl)}') |
| 799 | key_str := key.str() |
| 800 | if _ := tbl[key_str] { |
| 801 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 802 | ' key "${key}" is already initialized with a value. At "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 803 | } |
| 804 | tbl[key_str] = val |
| 805 | } |
| 806 | } |
| 807 | p.peek_for_correct_line_ending_or_fail()! |
| 808 | } |
| 809 | .eof { |
| 810 | break |
| 811 | } |
| 812 | else { |
| 813 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 814 | ' could not parse "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 815 | } |
| 816 | } |
| 817 | } |
| 818 | } |
| 819 | |
| 820 | // inline_table parses next tokens into a map of `ast.Value`s. |
| 821 | // The V map type is corresponding to a "table" in TOML. |
| 822 | pub fn (mut p Parser) inline_table(mut tbl map[string]ast.Value) ! { |
| 823 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing inline table into ${ptr_str(tbl)}...') |
| 824 | defer { p.value_is_immutable = true } |
| 825 | mut previous_token_was_value := false |
| 826 | for p.tok.kind != .eof { |
| 827 | p.next()! |
| 828 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing token "${p.tok.kind}"') |
| 829 | |
| 830 | if previous_token_was_value { |
| 831 | p.ignore_while(space_formatting_kinds()) |
| 832 | if p.tok.kind != .rcbr { |
| 833 | p.expect(.comma)! |
| 834 | } |
| 835 | previous_token_was_value = false |
| 836 | } |
| 837 | |
| 838 | match p.tok.kind { |
| 839 | .whitespace, .tab { |
| 840 | /* |
| 841 | if !p.scanner.config.tokenize_formatting { |
| 842 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'skipping "${p.tok.kind}" "${p.tok.lit}"') |
| 843 | continue |
| 844 | }*/ |
| 845 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 846 | 'skipping formatting "${p.tok.kind}" "${p.tok.lit}"') |
| 847 | continue |
| 848 | } |
| 849 | .comma { |
| 850 | p.ignore_while_peek(space_formatting_kinds()) |
| 851 | if p.peek_tok.kind in [.comma, .rcbr] { |
| 852 | p.next()! // Forward to the peek_tok |
| 853 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 854 | ' unexpected "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 855 | } |
| 856 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 857 | 'skipping comma table value separator "${p.tok.lit}"') |
| 858 | continue |
| 859 | } |
| 860 | .rcbr { |
| 861 | // '}' bracket |
| 862 | return |
| 863 | } |
| 864 | .bare, .quoted, .number, .minus, .underscore { |
| 865 | // Peek forward as far as we can skipping over space formatting tokens. |
| 866 | peek_tok, _ := p.peek_over(1, space_formatting_kinds())! |
| 867 | |
| 868 | if peek_tok.kind == .period { |
| 869 | dotted_key, val := p.dotted_key_value()! |
| 870 | |
| 871 | sub_table, key := p.sub_table_key(dotted_key) |
| 872 | mut t := p.find_in_table(mut tbl, sub_table)! |
| 873 | |
| 874 | // Disallow mutation of immutable values (inline tables) |
| 875 | if p.explicit_declared.has(dotted_key) { |
| 876 | left_most := DottedKey([dotted_key[0]]) |
| 877 | if t.len > 0 { |
| 878 | p.check_immutable(left_most)! |
| 879 | } |
| 880 | } |
| 881 | key_str := key.str() |
| 882 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 883 | 'inserting @6 "${key_str}" = ${val} into ${ptr_str(t)}') |
| 884 | unsafe { |
| 885 | if _ := t[key_str] { |
| 886 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 887 | ' key "${key_str}" is already initialized with a value. At "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 888 | } |
| 889 | t[key_str] = val |
| 890 | } |
| 891 | } else { |
| 892 | p.ignore_while(space_formatting_kinds()) |
| 893 | key, val := p.key_value()! |
| 894 | key_str := key.str() |
| 895 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 896 | 'inserting @5 "${key_str}" = ${val} into ${ptr_str(tbl)}') |
| 897 | if _ := tbl[key_str] { |
| 898 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 899 | ' key "${key_str}" is already initialized with a value. At "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 900 | } |
| 901 | tbl[key_str] = val |
| 902 | } |
| 903 | previous_token_was_value = true |
| 904 | } |
| 905 | else { |
| 906 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 907 | ' unexpected "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 908 | } |
| 909 | } |
| 910 | } |
| 911 | // Make sure the inline-table actually use the return at .rcbr match branch. |
| 912 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 913 | ' unexpected end of inline-table "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 914 | } |
| 915 | |
| 916 | // array_of_tables parses next tokens into an array of `ast.Value`s. |
| 917 | @[autofree_bug; manualfree] |
| 918 | pub fn (mut p Parser) array_of_tables(mut table map[string]ast.Value) ! { |
| 919 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 920 | 'parsing array of tables "${p.tok.kind}" "${p.tok.lit}"') |
| 921 | // NOTE this is starting to get ugly. TOML isn't simple at this point |
| 922 | p.check(.lsbr)! // '[' bracket |
| 923 | |
| 924 | // Allow [[ key]] |
| 925 | p.ignore_while(space_formatting_kinds()) |
| 926 | peek_tok, _ := p.peek_over(1, space_formatting_kinds())! |
| 927 | p.ignore_while(space_formatting_kinds()) |
| 928 | |
| 929 | // [[key.key]] horror |
| 930 | if peek_tok.kind == .period { |
| 931 | p.double_array_of_tables(mut table)! |
| 932 | return |
| 933 | } |
| 934 | |
| 935 | key := p.key()! |
| 936 | p.next()! |
| 937 | |
| 938 | // Allow [[key ]] |
| 939 | p.ignore_while(space_formatting_kinds()) |
| 940 | |
| 941 | p.check(.rsbr)! |
| 942 | p.peek_for_correct_line_ending_or_fail()! |
| 943 | p.expect(.rsbr)! |
| 944 | |
| 945 | p.ignore_while(all_formatting_kinds()) |
| 946 | |
| 947 | dotted_key := DottedKey([key.str()]) |
| 948 | dotted_key_str := dotted_key.str() |
| 949 | |
| 950 | // Disallow re-declaring the key |
| 951 | p.check_explicitly_declared(dotted_key)! |
| 952 | unsafe { |
| 953 | if val := table[dotted_key_str] { |
| 954 | if val is []ast.Value { |
| 955 | arr := &(table[dotted_key_str] as []ast.Value) |
| 956 | arr << p.array_of_tables_contents()! |
| 957 | table[dotted_key_str] = arr |
| 958 | } else { |
| 959 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 960 | ' table[${dotted_key_str}] is not an array. (excerpt): "...${p.excerpt()}..."') |
| 961 | } |
| 962 | } else { |
| 963 | table[dotted_key_str] = p.array_of_tables_contents()! |
| 964 | } |
| 965 | } |
| 966 | p.last_aot = dotted_key |
| 967 | |
| 968 | unsafe { |
| 969 | arr := &(table[p.last_aot.str()] as []ast.Value) |
| 970 | p.last_aot_index = arr.len - 1 |
| 971 | } |
| 972 | } |
| 973 | |
| 974 | // array_of_tables_contents parses next tokens into an array of `ast.Value`s. |
| 975 | @[autofree_bug; manualfree] |
| 976 | pub fn (mut p Parser) array_of_tables_contents() ![]ast.Value { |
| 977 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 978 | 'parsing contents from "${p.tok.kind}" "${p.tok.lit}"') |
| 979 | mut tbl := map[string]ast.Value{} |
| 980 | |
| 981 | p.table_contents(mut tbl)! |
| 982 | |
| 983 | mut arr := []ast.Value{} |
| 984 | arr << tbl |
| 985 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 986 | 'parsed array of tables ${ast.Value(arr)}. leaving at "${p.tok.kind}" "${p.tok.lit}"') |
| 987 | return arr |
| 988 | } |
| 989 | |
| 990 | // double_array_of_tables parses next tokens into an array of tables of arrays of `ast.Value`s... |
| 991 | @[autofree_bug; manualfree] |
| 992 | pub fn (mut p Parser) double_array_of_tables(mut table map[string]ast.Value) ! { |
| 993 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 994 | 'parsing nested array of tables "${p.tok.kind}" "${p.tok.lit}"') |
| 995 | |
| 996 | dotted_key := p.dotted_key()! |
| 997 | p.ignore_while(space_formatting_kinds()) |
| 998 | |
| 999 | p.check(.rsbr)! |
| 1000 | p.expect(.rsbr)! |
| 1001 | |
| 1002 | p.ignore_while(all_formatting_kinds()) |
| 1003 | |
| 1004 | p.check_explicitly_declared(dotted_key)! |
| 1005 | if is_all_tables(p.root_map, dotted_key) { |
| 1006 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1007 | ' key `${dotted_key.str()}` is already declared. Unexpected redeclaration at "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 1008 | } |
| 1009 | |
| 1010 | if !p.explicit_declared_array_of_tables.has(dotted_key) { |
| 1011 | p.explicit_declared_array_of_tables << dotted_key |
| 1012 | } |
| 1013 | |
| 1014 | first := DottedKey([dotted_key[0]]) // The array that holds the entries |
| 1015 | last := DottedKey([dotted_key[1]]) // The key the parsed array data should be added to |
| 1016 | |
| 1017 | mut t_arr := &[]ast.Value(unsafe { nil }) |
| 1018 | mut t_map := ast.Value(ast.Null{}) |
| 1019 | |
| 1020 | unsafe { |
| 1021 | if dotted_key.len == 2 { |
| 1022 | if table_first := table[first.str()] { |
| 1023 | if table_first is map[string]ast.Value { |
| 1024 | mut t := &(table_first as map[string]ast.Value) |
| 1025 | if val := t[last.str()] { |
| 1026 | if val is []ast.Value { |
| 1027 | mut arr := &val |
| 1028 | arr << p.array_of_tables_contents()! |
| 1029 | t[last.str()] = arr |
| 1030 | } else { |
| 1031 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1032 | ' t[${last.str()}] is not an array. (excerpt): "...${p.excerpt()}..."') |
| 1033 | } |
| 1034 | } else { |
| 1035 | t[last.str()] = p.array_of_tables_contents()! |
| 1036 | } |
| 1037 | p.last_aot.clear() |
| 1038 | p.last_aot_index = 0 |
| 1039 | return |
| 1040 | } |
| 1041 | } else { |
| 1042 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1043 | 'implicit allocation of map for `${first}` in dotted key `${dotted_key}`.') |
| 1044 | mut t := &map[string]ast.Value{} |
| 1045 | p.implicit_declared << first |
| 1046 | // NOTE: We register this implicit allocation also as *explicit* to be able to catch a special case like: |
| 1047 | // https://github.com/toml-lang/toml-test/blob/576db852/tests/invalid/table/array-implicit.toml |
| 1048 | // See also: undo_special_case_01 |
| 1049 | p.explicit_declared << first |
| 1050 | t[last.str()] = p.array_of_tables_contents()! |
| 1051 | table[first.str()] = ast.Value(t) |
| 1052 | p.last_aot.clear() |
| 1053 | p.last_aot_index = 0 |
| 1054 | return |
| 1055 | } |
| 1056 | } |
| 1057 | |
| 1058 | // NOTE this is starting to get EVEN uglier. TOML is not *at all* simple at this point... |
| 1059 | if first != p.last_aot { |
| 1060 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, '${first} != ${p.last_aot}') |
| 1061 | // Implicit allocation |
| 1062 | if p.last_aot.len == 0 { |
| 1063 | p.last_aot = first |
| 1064 | mut nm := &p.root_map |
| 1065 | if first.str() in table.keys() { |
| 1066 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1067 | 'adding to existing table entry at `${first}`.') |
| 1068 | table_first := table[first.str()] |
| 1069 | if table_first !is map[string]ast.Value { |
| 1070 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1071 | ' expected a table at "${first.str()}" but got "${table_first.type_name()}" instead. (excerpt): "...${p.excerpt()}..."') |
| 1072 | } |
| 1073 | nm = &(table_first as map[string]ast.Value) |
| 1074 | } else { |
| 1075 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1076 | 'implicit allocation of map for `${first}` in dotted key `${dotted_key}`.') |
| 1077 | nm = &map[string]ast.Value{} |
| 1078 | p.implicit_declared << first |
| 1079 | // NOTE: We register this implicit allocation also as *explicit* to be able to catch a special case like: |
| 1080 | // https://github.com/toml-lang/toml-test/blob/576db852/tests/invalid/table/array-implicit.toml |
| 1081 | // See also: undo_special_case_01 |
| 1082 | p.explicit_declared << first |
| 1083 | } |
| 1084 | |
| 1085 | nm[last.str()] = []ast.Value{} |
| 1086 | table[first.str()] = ast.Value(nm) |
| 1087 | |
| 1088 | t_arr = &(nm[last.str()] as []ast.Value) |
| 1089 | t_arr << p.array_of_tables_contents()! |
| 1090 | return |
| 1091 | } else { |
| 1092 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1093 | ' nested array of tables key "${first}" does not match "${p.last_aot}". (excerpt): "...${p.excerpt()}..."') |
| 1094 | } |
| 1095 | } |
| 1096 | |
| 1097 | array_of_tables := table[p.last_aot.str()] |
| 1098 | if first == p.last_aot { |
| 1099 | if array_of_tables is map[string]ast.Value { |
| 1100 | // NOTE: Here we "undo" the implicit-explicit special case declaration for: |
| 1101 | // https://github.com/toml-lang/toml-test/blob/576db852/tests/invalid/table/array-implicit.toml |
| 1102 | // ... to make the following test pass: |
| 1103 | // https://github.com/toml-lang/toml-test/blob/229ce2e/tests/valid/array/open-parent-table.toml |
| 1104 | p.undo_special_case_01(dotted_key) |
| 1105 | p.next()! |
| 1106 | return |
| 1107 | } |
| 1108 | } |
| 1109 | |
| 1110 | // Give a nicer error if the `as` cast below can not be done |
| 1111 | if array_of_tables !is []ast.Value { |
| 1112 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1113 | ' nested array of tables "${p.last_aot}" expected an array but got "${table[p.last_aot.str()].type_name()}". Re-definition is not allowed. (excerpt): "...${p.excerpt()}..."') |
| 1114 | } |
| 1115 | t_arr = &(array_of_tables as []ast.Value) |
| 1116 | t_map = ast.Value(map[string]ast.Value{}) |
| 1117 | if p.last_aot_index < t_arr.len { |
| 1118 | t_map = t_arr[p.last_aot_index] |
| 1119 | } |
| 1120 | |
| 1121 | if t_map !is map[string]ast.Value { |
| 1122 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1123 | ' expected a table but got "${t_map.type_name()}". (excerpt): "...${p.excerpt()}..."') |
| 1124 | } |
| 1125 | mut t := &(t_map as map[string]ast.Value) |
| 1126 | |
| 1127 | if val := t[last.str()] { |
| 1128 | if val is []ast.Value { |
| 1129 | mut arr := &val |
| 1130 | arr << p.double_array_of_tables_contents(dotted_key)! |
| 1131 | t[last.str()] = arr |
| 1132 | } else { |
| 1133 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1134 | ' t[${last.str()}] is not an array. (excerpt): "...${p.excerpt()}..."') |
| 1135 | } |
| 1136 | } else { |
| 1137 | t[last.str()] = p.double_array_of_tables_contents(dotted_key)! |
| 1138 | } |
| 1139 | if t_arr.len == 0 { |
| 1140 | t_arr << t |
| 1141 | p.last_aot_index = t_arr.len - 1 |
| 1142 | } |
| 1143 | } |
| 1144 | } |
| 1145 | |
| 1146 | // double_array_of_tables_contents parses next tokens into an array of `ast.Value`s. |
| 1147 | @[autofree_bug; manualfree] |
| 1148 | pub fn (mut p Parser) double_array_of_tables_contents(target_key DottedKey) ![]ast.Value { |
| 1149 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1150 | 'parsing contents from "${p.tok.kind}" "${p.tok.lit}"') |
| 1151 | mut tbl := map[string]ast.Value{} |
| 1152 | |
| 1153 | mut implicit_allocation_key := DottedKey([]string{}) |
| 1154 | mut peeked_over := 0 |
| 1155 | mut peek_tok := p.peek_tok |
| 1156 | |
| 1157 | for p.tok.kind != .eof { |
| 1158 | p.next()! |
| 1159 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing token "${p.tok.kind}"') |
| 1160 | p.ignore_while(all_formatting_kinds()) |
| 1161 | |
| 1162 | // Peek forward as far as we can skipping over space formatting tokens. |
| 1163 | peek_tok, peeked_over = p.peek_over(1, space_formatting_kinds())! |
| 1164 | // Peek for occurrence of `[[` |
| 1165 | if peek_tok.kind == .lsbr { |
| 1166 | peek_tok, peeked_over = p.peek_over(peeked_over + 1, space_formatting_kinds())! |
| 1167 | if peek_tok.kind == .lsbr { |
| 1168 | mut arr := []ast.Value{} |
| 1169 | arr << tbl |
| 1170 | return arr |
| 1171 | } |
| 1172 | } |
| 1173 | |
| 1174 | match p.tok.kind { |
| 1175 | .bare, .quoted, .number, .minus, .underscore { |
| 1176 | // Peek forward as far as we can skipping over space formatting tokens. |
| 1177 | peek_tok, _ = p.peek_over(1, space_formatting_kinds())! |
| 1178 | |
| 1179 | if peek_tok.kind == .period { |
| 1180 | mut dotted_key, val := p.dotted_key_value()! |
| 1181 | |
| 1182 | if implicit_allocation_key.len > 0 { |
| 1183 | dotted_key.insert(0, implicit_allocation_key) |
| 1184 | } |
| 1185 | sub_table, key := p.sub_table_key(dotted_key) |
| 1186 | |
| 1187 | mut t := p.find_in_table(mut tbl, sub_table)! |
| 1188 | unsafe { |
| 1189 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1190 | 'inserting @6 "${key}" = ${val} into ${ptr_str(t)}') |
| 1191 | t[key.str()] = val |
| 1192 | } |
| 1193 | } else { |
| 1194 | key, val := p.key_value()! |
| 1195 | |
| 1196 | mut t := unsafe { &tbl } |
| 1197 | if implicit_allocation_key.len > 0 { |
| 1198 | t = p.find_in_table(mut tbl, implicit_allocation_key)! |
| 1199 | } |
| 1200 | unsafe { |
| 1201 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1202 | 'inserting @7 "${key}" = ${val} into ${ptr_str(t)}') |
| 1203 | t[key.str()] = val |
| 1204 | } |
| 1205 | } |
| 1206 | } |
| 1207 | .lsbr { |
| 1208 | p.check(.lsbr)! // '[' bracket |
| 1209 | peek_tok = p.peek_tok |
| 1210 | |
| 1211 | // Allow `[ d.e.f]` |
| 1212 | p.ignore_while(space_formatting_kinds()) |
| 1213 | |
| 1214 | // Peek forward as far as we can skipping over space formatting tokens. |
| 1215 | peek_tok, _ = p.peek_over(1, space_formatting_kinds())! |
| 1216 | |
| 1217 | if peek_tok.kind == .period { |
| 1218 | // Parse `[d.e.f]` |
| 1219 | p.ignore_while(space_formatting_kinds()) |
| 1220 | dotted_key := p.dotted_key()! |
| 1221 | implicit_allocation_key = unsafe { dotted_key } |
| 1222 | if dotted_key.len > 2 { |
| 1223 | implicit_allocation_key = dotted_key[2..] |
| 1224 | } |
| 1225 | p.ignore_while(space_formatting_kinds()) |
| 1226 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1227 | 'keys are: dotted `${dotted_key}`, target `${target_key}`, implicit `${implicit_allocation_key}` at "${p.tok.kind}" "${p.tok.lit}"') |
| 1228 | p.expect(.rsbr)! |
| 1229 | p.peek_for_correct_line_ending_or_fail()! |
| 1230 | p.explicit_declared << dotted_key |
| 1231 | continue |
| 1232 | } else { |
| 1233 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1234 | ' could not parse "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 1235 | } |
| 1236 | } |
| 1237 | else { |
| 1238 | break |
| 1239 | } |
| 1240 | } |
| 1241 | } |
| 1242 | mut arr := []ast.Value{} |
| 1243 | arr << tbl |
| 1244 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1245 | 'parsed array of tables ${ast.Value(arr)}. leaving at "${p.tok.kind}" "${p.tok.lit}"') |
| 1246 | return arr |
| 1247 | } |
| 1248 | |
| 1249 | // array parses next tokens into an array of `ast.Value`s. |
| 1250 | pub fn (mut p Parser) array() ![]ast.Value { |
| 1251 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing array...') |
| 1252 | mut arr := []ast.Value{} |
| 1253 | p.expect(.lsbr)! // '[' bracket |
| 1254 | mut previous_token_was_value := false |
| 1255 | for p.tok.kind != .eof { |
| 1256 | p.next()! |
| 1257 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1258 | 'parsing token "${p.tok.kind}" "${p.tok.lit}"') |
| 1259 | |
| 1260 | if previous_token_was_value { |
| 1261 | p.ignore_while(all_formatting_kinds()) |
| 1262 | if p.tok.kind != .rsbr && p.tok.kind != .hash { |
| 1263 | p.expect(.comma)! |
| 1264 | } |
| 1265 | previous_token_was_value = false |
| 1266 | } else { |
| 1267 | if p.tok.kind == .comma { |
| 1268 | p.ignore_while_peek(space_formatting_kinds()) |
| 1269 | if p.peek_tok.kind == .rsbr { |
| 1270 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1271 | ' unexpected empty value in array "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 1272 | } |
| 1273 | } |
| 1274 | } |
| 1275 | p.ignore_while(all_formatting_kinds()) |
| 1276 | match p.tok.kind { |
| 1277 | .boolean { |
| 1278 | arr << ast.Value(p.boolean()!) |
| 1279 | previous_token_was_value = true |
| 1280 | } |
| 1281 | .comma { |
| 1282 | p.ignore_while_peek(space_formatting_kinds()) |
| 1283 | // Trailing commas before array close is allowed |
| 1284 | // so we do not do `if p.peek_tok.kind == .rsbr { ... }` |
| 1285 | |
| 1286 | // Check for known errors: |
| 1287 | if p.peek_tok.kind in [.comma, .bare] { |
| 1288 | p.next()! // Forward to the peek_tok |
| 1289 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1290 | ' unexpected "${p.tok.kind}" "${p.tok.lit}" at this (excerpt): "...${p.excerpt()}..."') |
| 1291 | } |
| 1292 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1293 | 'skipping comma table value separator "${p.tok.lit}"') |
| 1294 | continue |
| 1295 | } |
| 1296 | .eof { |
| 1297 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1298 | ' could not parse array. Reached EOF "${p.tok.kind}" "${p.tok.lit}" ("${p.tok.lit}") in this (excerpt): "...${p.excerpt()}..."') |
| 1299 | } |
| 1300 | .hash { |
| 1301 | c := p.comment() |
| 1302 | p.ast_root.comments << c |
| 1303 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'skipping comment "${c.text}"') |
| 1304 | } |
| 1305 | .lcbr { |
| 1306 | p.ignore_while(space_formatting_kinds()) |
| 1307 | mut t := map[string]ast.Value{} |
| 1308 | p.inline_table(mut t)! |
| 1309 | arr << ast.Value(t) |
| 1310 | previous_token_was_value = true |
| 1311 | } |
| 1312 | .number { |
| 1313 | val := p.number_or_date()! |
| 1314 | arr << val |
| 1315 | previous_token_was_value = true |
| 1316 | } |
| 1317 | .quoted { |
| 1318 | arr << ast.Value(p.quoted()) |
| 1319 | previous_token_was_value = true |
| 1320 | } |
| 1321 | .lsbr { |
| 1322 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1323 | 'parsing array in array "${p.tok.kind}" "${p.tok.lit}"') |
| 1324 | arr << ast.Value(p.array()!) |
| 1325 | previous_token_was_value = true |
| 1326 | } |
| 1327 | .rsbr { |
| 1328 | break |
| 1329 | } |
| 1330 | .bare { |
| 1331 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1332 | ' unexpected value "${p.tok.lit}". Array values should be quoted (with " or \') in this (excerpt): "...${p.excerpt()}..."') |
| 1333 | } |
| 1334 | else { |
| 1335 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1336 | ' unexpected token "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 1337 | } |
| 1338 | } |
| 1339 | } |
| 1340 | p.expect(.rsbr)! // ']' bracket |
| 1341 | $if debug { |
| 1342 | flat := arr.str().replace('\n', r'\n') |
| 1343 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1344 | 'parsed array: ${flat} . Currently @ token "${p.tok.kind}"') |
| 1345 | } |
| 1346 | return arr |
| 1347 | } |
| 1348 | |
| 1349 | // comment returns an `ast.Comment` type. |
| 1350 | pub fn (mut p Parser) comment() ast.Comment { |
| 1351 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed hash comment "#${p.tok.lit}"') |
| 1352 | return ast.Comment{ |
| 1353 | text: p.tok.lit |
| 1354 | pos: p.tok.pos() |
| 1355 | } |
| 1356 | } |
| 1357 | |
| 1358 | // key parse and returns an `ast.Key` type. |
| 1359 | // Keys are the token(s) appearing before an assignment operator (=). |
| 1360 | pub fn (mut p Parser) key() !ast.Key { |
| 1361 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing key from "${p.tok.lit}" ...') |
| 1362 | |
| 1363 | mut key := ast.Key(ast.Null{}) |
| 1364 | if p.tok.kind == .number { |
| 1365 | if p.peek_tok.kind == .minus { |
| 1366 | mut lits := p.tok.lit |
| 1367 | pos := p.tok.pos() |
| 1368 | for p.peek_tok.kind != .assign && p.peek_tok.kind != .period && p.peek_tok.kind != .rsbr { |
| 1369 | p.next()! |
| 1370 | if p.tok.kind !in space_formatting_kinds() { |
| 1371 | lits += p.tok.lit |
| 1372 | } |
| 1373 | } |
| 1374 | return ast.Key(ast.Bare{ |
| 1375 | text: lits |
| 1376 | pos: pos |
| 1377 | }) |
| 1378 | } |
| 1379 | num := p.number() |
| 1380 | // Handles if key is `1key` |
| 1381 | if p.peek_tok.kind in [.bare, .underscore, .minus] { |
| 1382 | bare := p.bare()! |
| 1383 | return bare |
| 1384 | } |
| 1385 | key = ast.Key(num) |
| 1386 | } else { |
| 1387 | key = match p.tok.kind { |
| 1388 | .bare, .underscore, .minus { |
| 1389 | ast.Key(p.bare()!) |
| 1390 | } |
| 1391 | .boolean { |
| 1392 | ast.Key(p.boolean()!) |
| 1393 | } |
| 1394 | .quoted { |
| 1395 | ast.Key(p.quoted()) |
| 1396 | } |
| 1397 | else { |
| 1398 | ast.Key(ast.Null{}) |
| 1399 | } |
| 1400 | } |
| 1401 | } |
| 1402 | |
| 1403 | // NOTE kept for eased debugging |
| 1404 | // util.printdbg(@MOD +'.' + @STRUCT + '.' + @FN, 'parsed key "${p.tok.lit}"') |
| 1405 | // panic(@MOD + '.' + @STRUCT + '.' + @FN + ' could not parse ${p.tok.kind} ("${p.tok.lit}") token \n${p.tok}') |
| 1406 | // return ast.Key(ast.Bare{}) |
| 1407 | |
| 1408 | if key is ast.Null { |
| 1409 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1410 | ' key expected .bare, .underscore, .number, .quoted or .boolean but got "${p.tok.kind}"') |
| 1411 | } |
| 1412 | |
| 1413 | // A few small exceptions that can't easily be done via `checker` or `decoder` *after* the |
| 1414 | // main table has been build since information like `is_multiline` is lost when using the key.text as a |
| 1415 | // V `map` key directly. |
| 1416 | mut decoded_key := key |
| 1417 | if key is ast.Quoted { |
| 1418 | if p.config.run_checks { |
| 1419 | quoted := key as ast.Quoted |
| 1420 | if quoted.is_multiline { |
| 1421 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1422 | ' multiline string as key is not allowed. (excerpt): "...${p.excerpt()}..."') |
| 1423 | } |
| 1424 | chckr := checker.Checker{ |
| 1425 | scanner: p.scanner |
| 1426 | } |
| 1427 | chckr.check_quoted(quoted)! |
| 1428 | } |
| 1429 | if p.config.decode_values { |
| 1430 | mut quoted := key as ast.Quoted |
| 1431 | decoder.decode_quoted_escapes(mut quoted)! |
| 1432 | decoded_key = ast.Key(quoted) |
| 1433 | } |
| 1434 | } |
| 1435 | |
| 1436 | return decoded_key |
| 1437 | } |
| 1438 | |
| 1439 | // key_value parse and returns a pair `ast.Key` and `ast.Value` type. |
| 1440 | // see also `key()` and `value()` |
| 1441 | pub fn (mut p Parser) key_value() !(ast.Key, ast.Value) { |
| 1442 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing key value pair...') |
| 1443 | key := p.key()! |
| 1444 | dotted_key := DottedKey([key.str()]) |
| 1445 | p.explicit_declared << p.build_abs_dotted_key(dotted_key) |
| 1446 | p.next()! |
| 1447 | p.ignore_while(space_formatting_kinds()) |
| 1448 | p.check(.assign)! // Assignment operator |
| 1449 | p.ignore_while(space_formatting_kinds()) |
| 1450 | value := p.value()! |
| 1451 | if p.value_is_immutable { |
| 1452 | if !p.immutable.has(dotted_key) { |
| 1453 | p.immutable << p.build_abs_dotted_key(dotted_key) // Mark the key we are assigning to as immutable |
| 1454 | } |
| 1455 | p.value_is_immutable = false |
| 1456 | } |
| 1457 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed key value pair. `${key} = ${value}`') |
| 1458 | return key, value |
| 1459 | } |
| 1460 | |
| 1461 | // dotted_key_value parse and returns a pair `DottedKey` and `ast.Value` type. |
| 1462 | // see also `key()` and `value()` |
| 1463 | pub fn (mut p Parser) dotted_key_value() !(DottedKey, ast.Value) { |
| 1464 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsing dotted key value pair...') |
| 1465 | p.ignore_while(space_formatting_kinds()) |
| 1466 | dotted_key := p.dotted_key()! |
| 1467 | p.explicit_declared << p.build_abs_dotted_key(dotted_key) |
| 1468 | p.ignore_while(space_formatting_kinds()) |
| 1469 | p.check(.assign)! |
| 1470 | p.ignore_while(space_formatting_kinds()) |
| 1471 | value := p.value()! |
| 1472 | if p.value_is_immutable { |
| 1473 | if !p.immutable.has(dotted_key) { |
| 1474 | p.immutable << p.build_abs_dotted_key(dotted_key) // Mark the key we are assigning to as immutable |
| 1475 | } |
| 1476 | p.value_is_immutable = false |
| 1477 | } |
| 1478 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1479 | 'parsed dotted key value pair `${dotted_key} = ${value}`...') |
| 1480 | |
| 1481 | return dotted_key, value |
| 1482 | } |
| 1483 | |
| 1484 | // value parse and returns an `ast.Value` type. |
| 1485 | // values are the token(s) appearing after an assignment operator (=). |
| 1486 | pub fn (mut p Parser) value() !ast.Value { |
| 1487 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, |
| 1488 | 'parsing value from token "${p.tok.kind}" "${p.tok.lit}"...') |
| 1489 | mut value := ast.Value(ast.Null{}) |
| 1490 | if p.tok.kind == .number { |
| 1491 | number_or_date := p.number_or_date()! |
| 1492 | value = number_or_date |
| 1493 | } else { |
| 1494 | value = match p.tok.kind { |
| 1495 | .quoted { |
| 1496 | ast.Value(p.quoted()) |
| 1497 | } |
| 1498 | .boolean { |
| 1499 | ast.Value(p.boolean()!) |
| 1500 | } |
| 1501 | .lsbr { |
| 1502 | ast.Value(p.array()!) |
| 1503 | } |
| 1504 | .lcbr { |
| 1505 | p.ignore_while(space_formatting_kinds()) |
| 1506 | mut t := map[string]ast.Value{} |
| 1507 | p.inline_table(mut t)! |
| 1508 | ast.Value(t) |
| 1509 | } |
| 1510 | else { |
| 1511 | ast.Value(ast.Null{}) |
| 1512 | } |
| 1513 | } |
| 1514 | |
| 1515 | if value is ast.Null { |
| 1516 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1517 | ' value expected .boolean, .quoted, .lsbr, .lcbr or .number got "${p.tok.kind}" "${p.tok.lit}" in this (excerpt): "...${p.excerpt()}..."') |
| 1518 | } |
| 1519 | } |
| 1520 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed "${p.tok.kind}" as value ${value}') |
| 1521 | return value |
| 1522 | } |
| 1523 | |
| 1524 | // number_or_date parse and returns an `ast.Value` type as |
| 1525 | // one of [`ast.Date`, `ast.Time`, `ast.DateTime`, `ast.Number`] |
| 1526 | pub fn (mut p Parser) number_or_date() !ast.Value { |
| 1527 | // Handle Date/Time |
| 1528 | if p.peek_tok.kind == .minus || p.peek_tok.kind == .colon { |
| 1529 | date_time_type := p.date_time()! |
| 1530 | match date_time_type { |
| 1531 | ast.Date { |
| 1532 | return ast.Value(date_time_type) |
| 1533 | } |
| 1534 | ast.Time { |
| 1535 | return ast.Value(date_time_type) |
| 1536 | } |
| 1537 | ast.DateTime { |
| 1538 | return ast.Value(date_time_type) |
| 1539 | } |
| 1540 | } |
| 1541 | } |
| 1542 | return ast.Value(p.number()) |
| 1543 | } |
| 1544 | |
| 1545 | // bare parse and returns an `ast.Bare` type. |
| 1546 | pub fn (mut p Parser) bare() !ast.Bare { |
| 1547 | mut lits := p.tok.lit |
| 1548 | pos := p.tok.pos() |
| 1549 | for p.peek_tok.kind != .assign && p.peek_tok.kind != .period && p.peek_tok.kind != .rsbr |
| 1550 | && p.peek_tok.kind !in space_formatting_kinds() { |
| 1551 | p.next()! |
| 1552 | if p.tok.kind == .bare || p.tok.kind == .minus || p.tok.kind == .underscore { |
| 1553 | lits += p.tok.lit |
| 1554 | continue |
| 1555 | } |
| 1556 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1557 | ' bare key expected .bare, .minus, or .underscore but got "${p.tok.kind}"') |
| 1558 | } |
| 1559 | return ast.Bare{ |
| 1560 | text: lits |
| 1561 | pos: pos |
| 1562 | } |
| 1563 | } |
| 1564 | |
| 1565 | // quoted parse and returns an `ast.Quoted` type. |
| 1566 | pub fn (mut p Parser) quoted() ast.Quoted { |
| 1567 | // To get more info about the quote type and enable better checking, |
| 1568 | // the scanner is returning the literal *with* single- or double-quotes. |
| 1569 | mut quote := p.tok.lit[0] |
| 1570 | is_multiline := p.tok.lit.len >= 6 && p.tok.lit[1] == quote && p.tok.lit[2] == quote |
| 1571 | mut lit := p.tok.lit[1..p.tok.lit.len - 1] |
| 1572 | if is_multiline { |
| 1573 | lit = p.tok.lit[3..p.tok.lit.len - 3] |
| 1574 | // From https://toml.io/en/v1.0.0#string |
| 1575 | // "Multi-line literal strings [...] A newline immediately following the opening |
| 1576 | // delimiter will be trimmed. All other content between the delimiters |
| 1577 | // is interpreted as-is without modification." |
| 1578 | if lit.len > 0 && lit[0] == `\n` { |
| 1579 | lit = lit[1..] |
| 1580 | } |
| 1581 | } |
| 1582 | return ast.Quoted{ |
| 1583 | text: lit |
| 1584 | pos: p.tok.pos() |
| 1585 | quote: quote |
| 1586 | is_multiline: is_multiline |
| 1587 | } |
| 1588 | } |
| 1589 | |
| 1590 | // boolean parse and returns an `ast.Bool` type. |
| 1591 | pub fn (mut p Parser) boolean() !ast.Bool { |
| 1592 | if p.tok.lit !in ['true', 'false'] { |
| 1593 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1594 | ' expected literal to be either `true` or `false` got "${p.tok.kind}"') |
| 1595 | } |
| 1596 | return ast.Bool{ |
| 1597 | text: p.tok.lit |
| 1598 | pos: p.tok.pos() |
| 1599 | } |
| 1600 | } |
| 1601 | |
| 1602 | // number parse and returns an `ast.Number` type. |
| 1603 | pub fn (mut p Parser) number() ast.Number { |
| 1604 | return ast.Number{ |
| 1605 | text: p.tok.lit |
| 1606 | pos: p.tok.pos() |
| 1607 | } |
| 1608 | } |
| 1609 | |
| 1610 | // date_time parses dates and time in RFC 3339 format. |
| 1611 | // https://datatracker.ietf.org/doc/html/rfc3339 |
| 1612 | pub fn (mut p Parser) date_time() !ast.DateTimeType { |
| 1613 | // Date and/or Time |
| 1614 | mut lit := '' |
| 1615 | pos := p.tok.pos() |
| 1616 | mut date := ast.Date{} |
| 1617 | mut time := ast.Time{} |
| 1618 | |
| 1619 | if p.peek_tok.kind == .minus { |
| 1620 | date = p.date()! |
| 1621 | lit += date.text |
| 1622 | // Look for any THH:MM:SS or <space>HH:MM:SS |
| 1623 | if (p.peek_tok.kind == .bare && (p.peek_tok.lit.starts_with('T') |
| 1624 | || p.peek_tok.lit.starts_with('t'))) || p.peek_tok.kind == .whitespace { |
| 1625 | p.next()! // Advance to token with Txx or whitespace special case |
| 1626 | if p.tok.lit.starts_with('T') || p.tok.lit.starts_with('t') { |
| 1627 | lit += p.tok.lit[0].ascii_str() //'T' or 't' |
| 1628 | } else { |
| 1629 | peek := p.peek(0)! |
| 1630 | if peek.kind != .number { |
| 1631 | // return early as date for strings yyyy-mm-dd_X... (_ is space, X is not numeric) |
| 1632 | return ast.Date{ |
| 1633 | text: lit |
| 1634 | pos: pos |
| 1635 | } |
| 1636 | } |
| 1637 | lit += p.tok.lit |
| 1638 | p.next()! |
| 1639 | } |
| 1640 | time = p.time()! |
| 1641 | lit += time.text |
| 1642 | |
| 1643 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed date-time: "${lit}"') |
| 1644 | return ast.DateTime{ |
| 1645 | text: lit |
| 1646 | pos: pos |
| 1647 | date: date |
| 1648 | time: time |
| 1649 | } |
| 1650 | } |
| 1651 | } else if p.peek_tok.kind == .colon { |
| 1652 | time = p.time()! |
| 1653 | return time |
| 1654 | } |
| 1655 | |
| 1656 | return ast.Date{ |
| 1657 | text: lit |
| 1658 | pos: pos |
| 1659 | } |
| 1660 | } |
| 1661 | |
| 1662 | // date parse and returns an `ast.Date` type. |
| 1663 | pub fn (mut p Parser) date() !ast.Date { |
| 1664 | // Date |
| 1665 | mut lit := p.tok.lit |
| 1666 | pos := p.tok.pos() |
| 1667 | |
| 1668 | p.check(.number)! |
| 1669 | lit += p.tok.lit |
| 1670 | p.check(.minus)! |
| 1671 | lit += p.tok.lit |
| 1672 | p.check(.number)! |
| 1673 | lit += p.tok.lit |
| 1674 | p.check(.minus)! |
| 1675 | lit += p.tok.lit |
| 1676 | p.expect(.number)! |
| 1677 | |
| 1678 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed date: "${lit}"') |
| 1679 | return ast.Date{ |
| 1680 | text: lit |
| 1681 | pos: pos |
| 1682 | } |
| 1683 | } |
| 1684 | |
| 1685 | // time parse and returns an `ast.Time` type. |
| 1686 | pub fn (mut p Parser) time() !ast.Time { |
| 1687 | // Time |
| 1688 | mut lit := p.tok.lit |
| 1689 | pos := p.tok.pos() |
| 1690 | |
| 1691 | if p.is_at(.bare) && (lit.starts_with('T') || lit.starts_with('t')) { |
| 1692 | if p.tok.lit.starts_with('T') { |
| 1693 | lit = lit.all_after('T') |
| 1694 | } else if p.tok.lit.starts_with('t') { |
| 1695 | lit = lit.all_after('t') |
| 1696 | } |
| 1697 | p.next()! |
| 1698 | } else { |
| 1699 | p.check(.number)! |
| 1700 | } |
| 1701 | lit += p.tok.lit |
| 1702 | p.check(.colon)! |
| 1703 | lit += p.tok.lit |
| 1704 | p.check(.number)! |
| 1705 | lit += p.tok.lit |
| 1706 | // NOTE: TOML v1.1.0 have optional seconds |
| 1707 | // if p.peek_tok.kind == .colon { |
| 1708 | p.check(.colon)! |
| 1709 | lit += p.tok.lit |
| 1710 | p.expect(.number)! |
| 1711 | //} |
| 1712 | |
| 1713 | // Optional milliseconds |
| 1714 | if p.peek_tok.kind == .period { |
| 1715 | p.next()! |
| 1716 | lit += p.tok.lit // lit += '.' |
| 1717 | p.check(.period)! |
| 1718 | lit += p.tok.lit |
| 1719 | p.expect(.number)! |
| 1720 | } |
| 1721 | |
| 1722 | if !lit[lit.len - 1].is_digit() { |
| 1723 | return error(@MOD + '.' + @STRUCT + '.' + @FN + |
| 1724 | ' expected a number as last occurrence in "${lit}" got "${lit[lit.len - |
| 1725 | 1].ascii_str()}"') |
| 1726 | } |
| 1727 | |
| 1728 | // Parse offset |
| 1729 | if p.peek_tok.kind == .minus || p.peek_tok.kind == .plus { |
| 1730 | p.next()! |
| 1731 | lit += p.tok.lit // lit += '-' |
| 1732 | p.check_one_of([.minus, .plus])! |
| 1733 | lit += p.tok.lit |
| 1734 | p.check(.number)! |
| 1735 | lit += p.tok.lit |
| 1736 | p.check(.colon)! |
| 1737 | lit += p.tok.lit |
| 1738 | p.expect(.number)! |
| 1739 | } else if p.peek_tok.kind == .bare && (p.peek_tok.lit == 'Z' || p.peek_tok.lit == 'z') { |
| 1740 | p.next()! |
| 1741 | lit += p.tok.lit |
| 1742 | p.expect(.bare)! |
| 1743 | } |
| 1744 | |
| 1745 | util.printdbg(@MOD + '.' + @STRUCT + '.' + @FN, 'parsed time: "${lit}"') |
| 1746 | return ast.Time{ |
| 1747 | text: lit |
| 1748 | pos: pos |
| 1749 | } |
| 1750 | } |
| 1751 | |
| 1752 | // undo_special_case_01 reverts an operation needed for a few special case / edge case tests to pass. |
| 1753 | // See: |
| 1754 | // https://github.com/toml-lang/toml-test/blob/576db852/tests/invalid/table/array-implicit.toml |
| 1755 | // https://github.com/toml-lang/toml-test/blob/229ce2e/tests/valid/table/array-implicit-and-explicit-after.toml |
| 1756 | // https://github.com/toml-lang/toml-test/blob/229ce2e/tests/valid/array/open-parent-table.toml |
| 1757 | pub fn (mut p Parser) undo_special_case_01(dotted_key DottedKey) { |
| 1758 | exd_i := p.explicit_declared.index(dotted_key) |
| 1759 | if exd_i > -1 { |
| 1760 | p.explicit_declared.delete(exd_i) |
| 1761 | p.last_aot.clear() |
| 1762 | } |
| 1763 | } |
| 1764 | |
| 1765 | // eof returns an `ast.EOF` type. |
| 1766 | pub fn (mut p Parser) eof() ast.EOF { |
| 1767 | return ast.EOF{ |
| 1768 | pos: p.tok.pos() |
| 1769 | } |
| 1770 | } |
| 1771 | |