lexi.c revision 1.87
1 /* $NetBSD: lexi.c,v 1.87 2021/10/08 23:55:44 rillig Exp $ */ 2 3 /*- 4 * SPDX-License-Identifier: BSD-4-Clause 5 * 6 * Copyright (c) 1985 Sun Microsystems, Inc. 7 * Copyright (c) 1980, 1993 8 * The Regents of the University of California. All rights reserved. 9 * All rights reserved. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. All advertising materials mentioning features or use of this software 20 * must display the following acknowledgement: 21 * This product includes software developed by the University of 22 * California, Berkeley and its contributors. 23 * 4. Neither the name of the University nor the names of its contributors 24 * may be used to endorse or promote products derived from this software 25 * without specific prior written permission. 26 * 27 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 29 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 30 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 31 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 32 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 33 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 34 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 35 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 36 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 37 * SUCH DAMAGE. 38 */ 39 40 #if 0 41 static char sccsid[] = "@(#)lexi.c 8.1 (Berkeley) 6/6/93"; 42 #endif 43 44 #include <sys/cdefs.h> 45 #if defined(__NetBSD__) 46 __RCSID("$NetBSD: lexi.c,v 1.87 2021/10/08 23:55:44 rillig Exp $"); 47 #elif defined(__FreeBSD__) 48 __FBSDID("$FreeBSD: head/usr.bin/indent/lexi.c 337862 2018-08-15 18:19:45Z pstef $"); 49 #endif 50 51 #include <assert.h> 52 #include <stdio.h> 53 #include <ctype.h> 54 #include <stdlib.h> 55 #include <string.h> 56 #include <sys/param.h> 57 58 #include "indent.h" 59 60 /* must be sorted alphabetically, is used in binary search */ 61 static const struct keyword { 62 const char *name; 63 enum keyword_kind kind; 64 } keywords[] = { 65 {"_Bool", kw_type}, 66 {"_Complex", kw_type}, 67 {"_Imaginary", kw_type}, 68 {"auto", kw_storage_class}, 69 {"bool", kw_type}, 70 {"break", kw_jump}, 71 {"case", kw_case_or_default}, 72 {"char", kw_type}, 73 {"complex", kw_type}, 74 {"const", kw_type}, 75 {"continue", kw_jump}, 76 {"default", kw_case_or_default}, 77 {"do", kw_do_or_else}, 78 {"double", kw_type}, 79 {"else", kw_do_or_else}, 80 {"enum", kw_struct_or_union_or_enum}, 81 {"extern", kw_storage_class}, 82 {"float", kw_type}, 83 {"for", kw_for_or_if_or_while}, 84 {"goto", kw_jump}, 85 {"if", kw_for_or_if_or_while}, 86 {"imaginary", kw_type}, 87 {"inline", kw_inline_or_restrict}, 88 {"int", kw_type}, 89 {"long", kw_type}, 90 {"offsetof", kw_offsetof}, 91 {"register", kw_storage_class}, 92 {"restrict", kw_inline_or_restrict}, 93 {"return", kw_jump}, 94 {"short", kw_type}, 95 {"signed", kw_type}, 96 {"sizeof", kw_sizeof}, 97 {"static", kw_storage_class}, 98 {"struct", kw_struct_or_union_or_enum}, 99 {"switch", kw_switch}, 100 {"typedef", kw_typedef}, 101 {"union", kw_struct_or_union_or_enum}, 102 {"unsigned", kw_type}, 103 {"void", kw_type}, 104 {"volatile", kw_type}, 105 {"while", kw_for_or_if_or_while} 106 }; 107 108 static struct { 109 const char **items; 110 unsigned int len; 111 unsigned int cap; 112 } typenames; 113 114 /* 115 * The transition table below was rewritten by hand from lx's output, given 116 * the following definitions. lx is Katherine Flavel's lexer generator. 117 * 118 * O = /[0-7]/; D = /[0-9]/; NZ = /[1-9]/; 119 * H = /[a-f0-9]/i; B = /[0-1]/; HP = /0x/i; 120 * BP = /0b/i; E = /e[+\-]?/i D+; P = /p[+\-]?/i D+; 121 * FS = /[fl]/i; IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?; 122 * 123 * D+ E FS? -> $float; 124 * D* "." D+ E? FS? -> $float; 125 * D+ "." E? FS? -> $float; HP H+ IS? -> $int; 126 * HP H+ P FS? -> $float; NZ D* IS? -> $int; 127 * HP H* "." H+ P FS? -> $float; "0" O* IS? -> $int; 128 * HP H+ "." P FS -> $float; BP B+ IS? -> $int; 129 */ 130 /* INDENT OFF */ 131 static const unsigned char lex_number_state[][26] = { 132 /* examples: 133 00 134 s 0xx 135 t 00xaa 136 a 11 101100xxa.. 137 r 11ee0001101lbuuxx.a.pp 138 t.01.e+008bLuxll0Ll.aa.p+0 139 states: ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 140 [0] = "uuiifuufiuuiiuiiiiiuiuuuuu", /* (other) */ 141 [1] = "CEIDEHHHIJQ U Q VUVVZZZ", /* 0 */ 142 [2] = "DEIDEHHHIJQ U Q VUVVZZZ", /* 1 */ 143 [3] = "DEIDEHHHIJ U VUVVZZZ", /* 2 3 4 5 6 7 */ 144 [4] = "DEJDEHHHJJ U VUVVZZZ", /* 8 9 */ 145 [5] = " U VUVV ", /* A a C c D d */ 146 [6] = " K U VUVV ", /* B b */ 147 [7] = " FFF FF U VUVV ", /* E e */ 148 [8] = " f f U VUVV f", /* F f */ 149 [9] = " LLf fL PR Li L f", /* L */ 150 [10] = " OOf fO S P O i O f", /* l */ 151 [11] = " FFX ", /* P p */ 152 [12] = " MM M i iiM M ", /* U u */ 153 [13] = " N ", /* X x */ 154 [14] = " G Y ", /* + - */ 155 [15] = "B EE EE T W ", /* . */ 156 /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 157 }; 158 /* INDENT ON */ 159 160 static const uint8_t lex_number_row[] = { 161 ['0'] = 1, 162 ['1'] = 2, 163 ['2'] = 3, ['3'] = 3, ['4'] = 3, ['5'] = 3, ['6'] = 3, ['7'] = 3, 164 ['8'] = 4, ['9'] = 4, 165 ['A'] = 5, ['a'] = 5, ['C'] = 5, ['c'] = 5, ['D'] = 5, ['d'] = 5, 166 ['B'] = 6, ['b'] = 6, 167 ['E'] = 7, ['e'] = 7, 168 ['F'] = 8, ['f'] = 8, 169 ['L'] = 9, 170 ['l'] = 10, 171 ['P'] = 11, ['p'] = 11, 172 ['U'] = 12, ['u'] = 12, 173 ['X'] = 13, ['x'] = 13, 174 ['+'] = 14, ['-'] = 14, 175 ['.'] = 15, 176 }; 177 178 static char 179 inbuf_peek(void) 180 { 181 return *inp.s; 182 } 183 184 void 185 inbuf_skip(void) 186 { 187 inp.s++; 188 if (inp.s >= inp.e) 189 inbuf_read_line(); 190 } 191 192 char 193 inbuf_next(void) 194 { 195 char ch = inbuf_peek(); 196 inbuf_skip(); 197 return ch; 198 } 199 200 static void 201 check_size_token(size_t desired_size) 202 { 203 if (token.e + desired_size >= token.l) 204 buf_expand(&token, desired_size); 205 } 206 207 static void 208 token_add_char(char ch) 209 { 210 check_size_token(1); 211 *token.e++ = ch; 212 } 213 214 static int 215 cmp_keyword_by_name(const void *key, const void *elem) 216 { 217 return strcmp(key, ((const struct keyword *)elem)->name); 218 } 219 220 #ifdef debug 221 const char * 222 token_type_name(token_type ttype) 223 { 224 static const char *const name[] = { 225 "end_of_file", "newline", "lparen_or_lbracket", "rparen_or_rbracket", 226 "unary_op", "binary_op", "postfix_op", "question", 227 "case_label", "colon", 228 "semicolon", "lbrace", "rbrace", "ident", "comma", 229 "comment", "switch_expr", "preprocessing", "form_feed", "decl", 230 "keyword_for_if_while", "keyword_do_else", 231 "if_expr", "while_expr", "for_exprs", 232 "stmt", "stmt_list", "keyword_else", "keyword_do", "do_stmt", 233 "if_expr_stmt", "if_expr_stmt_else", "period", "string_prefix", 234 "storage_class", "funcname", "type_def", "keyword_struct_union_enum" 235 }; 236 237 assert(0 <= ttype && ttype < nitems(name)); 238 239 return name[ttype]; 240 } 241 242 static void 243 debug_print_buf(const char *name, const struct buffer *buf) 244 { 245 if (buf->s < buf->e) { 246 debug_printf(" %s ", name); 247 debug_vis_range("\"", buf->s, buf->e, "\""); 248 } 249 } 250 251 static token_type 252 lexi_end(token_type ttype) 253 { 254 debug_printf("in line %d, lexi returns '%s'", 255 line_no, token_type_name(ttype)); 256 debug_print_buf("token", &token); 257 debug_print_buf("label", &lab); 258 debug_print_buf("code", &code); 259 debug_print_buf("comment", &com); 260 debug_printf("\n"); 261 262 return ttype; 263 } 264 #else 265 #define lexi_end(tk) (tk) 266 #endif 267 268 static void 269 lex_number(void) 270 { 271 for (uint8_t s = 'A'; s != 'f' && s != 'i' && s != 'u';) { 272 uint8_t ch = (uint8_t)*inp.s; 273 if (ch >= nitems(lex_number_row) || lex_number_row[ch] == 0) 274 break; 275 276 uint8_t row = lex_number_row[ch]; 277 if (lex_number_state[row][s - 'A'] == ' ') { 278 /*- 279 * lex_number_state[0][s - 'A'] now indicates the type: 280 * f = floating, i = integer, u = unknown 281 */ 282 break; 283 } 284 285 s = lex_number_state[row][s - 'A']; 286 token_add_char(inbuf_next()); 287 } 288 } 289 290 static void 291 lex_word(void) 292 { 293 while (isalnum((unsigned char)*inp.s) || 294 *inp.s == '\\' || 295 *inp.s == '_' || *inp.s == '$') { 296 297 if (*inp.s == '\\') { 298 if (inp.s[1] == '\n') { 299 inp.s += 2; 300 if (inp.s >= inp.e) 301 inbuf_read_line(); 302 } else 303 break; 304 } 305 306 token_add_char(inbuf_next()); 307 } 308 } 309 310 static void 311 lex_char_or_string(void) 312 { 313 for (char delim = *token.s;;) { 314 if (*inp.s == '\n') { 315 diag(1, "Unterminated literal"); 316 return; 317 } 318 319 token_add_char(inbuf_next()); 320 if (token.e[-1] == delim) 321 return; 322 323 if (token.e[-1] == '\\') { 324 if (*inp.s == '\n') 325 ++line_no; 326 token_add_char(inbuf_next()); 327 } 328 } 329 } 330 331 /* Guess whether the current token is a declared type. */ 332 static bool 333 probably_typename(const struct parser_state *state) 334 { 335 if (state->p_l_follow != 0) 336 return false; 337 if (state->block_init || state->in_stmt) 338 return false; 339 if (inp.s[0] == '*' && inp.s[1] != '=') 340 goto maybe; 341 if (isalpha((unsigned char)*inp.s)) 342 goto maybe; 343 return false; 344 maybe: 345 return state->last_token == semicolon || 346 state->last_token == lbrace || 347 state->last_token == rbrace; 348 } 349 350 static int 351 bsearch_typenames(const char *key) 352 { 353 const char **arr = typenames.items; 354 int lo = 0; 355 int hi = (int)typenames.len - 1; 356 357 while (lo <= hi) { 358 int mid = (int)((unsigned)(lo + hi) >> 1); 359 int cmp = strcmp(arr[mid], key); 360 if (cmp < 0) 361 lo = mid + 1; 362 else if (cmp > 0) 363 hi = mid - 1; 364 else 365 return mid; 366 } 367 return -(lo + 1); 368 } 369 370 static bool 371 is_typename(void) 372 { 373 if (opt.auto_typedefs && 374 token.e - token.s >= 2 && memcmp(token.e - 2, "_t", 2) == 0) 375 return true; 376 377 return bsearch_typenames(token.s) >= 0; 378 } 379 380 /* Reads the next token, placing it in the global variable "token". */ 381 token_type 382 lexi(struct parser_state *state) 383 { 384 bool unary_delim; /* whether the current token forces a 385 * following operator to be unary */ 386 token_type ttype; 387 388 token.e = token.s; /* point to start of place to save token */ 389 unary_delim = false; 390 state->col_1 = state->last_nl; /* tell world that this token started 391 * in column 1 iff the last thing 392 * scanned was a newline */ 393 state->last_nl = false; 394 395 while (is_hspace(*inp.s)) { 396 state->col_1 = false; 397 inbuf_skip(); 398 } 399 400 /* Scan an alphanumeric token */ 401 if (isalnum((unsigned char)*inp.s) || 402 *inp.s == '_' || *inp.s == '$' || 403 (inp.s[0] == '.' && isdigit((unsigned char)inp.s[1]))) { 404 struct keyword *kw; 405 406 if (isdigit((unsigned char)*inp.s) || 407 (inp.s[0] == '.' && isdigit((unsigned char)inp.s[1]))) { 408 lex_number(); 409 } else { 410 lex_word(); 411 } 412 *token.e = '\0'; 413 414 if (token.s[0] == 'L' && token.s[1] == '\0' && 415 (*inp.s == '"' || *inp.s == '\'')) 416 return lexi_end(string_prefix); 417 418 while (is_hspace(inbuf_peek())) 419 inbuf_skip(); 420 state->keyword = kw_0; 421 422 if (state->last_token == keyword_struct_union_enum && 423 state->p_l_follow == 0) { 424 state->last_u_d = true; 425 return lexi_end(decl); 426 } 427 /* 428 * Operator after identifier is binary unless last token was 'struct' 429 */ 430 state->last_u_d = (state->last_token == keyword_struct_union_enum); 431 432 kw = bsearch(token.s, keywords, nitems(keywords), 433 sizeof(keywords[0]), cmp_keyword_by_name); 434 if (kw == NULL) { 435 if (is_typename()) { 436 state->keyword = kw_type; 437 state->last_u_d = true; 438 goto found_typename; 439 } 440 441 } else { /* we have a keyword */ 442 state->keyword = kw->kind; 443 state->last_u_d = true; 444 445 switch (kw->kind) { 446 case kw_switch: 447 return lexi_end(switch_expr); 448 449 case kw_case_or_default: 450 return lexi_end(case_label); 451 452 case kw_struct_or_union_or_enum: 453 case kw_type: 454 found_typename: 455 if (state->p_l_follow != 0) { 456 /* inside parens: cast, param list, offsetof or sizeof */ 457 state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask; 458 } 459 if (state->last_token == period || state->last_token == unary_op) { 460 state->keyword = kw_0; 461 break; 462 } 463 if (kw != NULL && kw->kind == kw_struct_or_union_or_enum) 464 return lexi_end(keyword_struct_union_enum); 465 if (state->p_l_follow != 0) 466 break; 467 return lexi_end(decl); 468 469 case kw_for_or_if_or_while: 470 return lexi_end(keyword_for_if_while); 471 472 case kw_do_or_else: 473 return lexi_end(keyword_do_else); 474 475 case kw_storage_class: 476 return lexi_end(storage_class); 477 478 case kw_typedef: 479 return lexi_end(type_def); 480 481 default: /* all others are treated like any other 482 * identifier */ 483 return lexi_end(ident); 484 } /* end of switch */ 485 } /* end of if (found_it) */ 486 487 if (*inp.s == '(' && state->tos <= 1 && state->ind_level == 0 && 488 !state->in_parameter_declaration && !state->block_init) { 489 490 for (const char *p = inp.s; p < inp.e;) 491 if (*p++ == ')' && (*p == ';' || *p == ',')) 492 goto not_proc; 493 494 strncpy(state->procname, token.s, sizeof state->procname - 1); 495 if (state->in_decl) 496 state->in_parameter_declaration = true; 497 return lexi_end(funcname); 498 not_proc:; 499 500 } else if (probably_typename(state)) { 501 state->keyword = kw_type; 502 state->last_u_d = true; 503 return lexi_end(decl); 504 } 505 506 if (state->last_token == decl) /* if this is a declared variable, 507 * then following sign is unary */ 508 state->last_u_d = true; /* will make "int a -1" work */ 509 510 return lexi_end(ident); /* the ident is not in the list */ 511 } /* end of processing for alphanum character */ 512 513 /* Scan a non-alphanumeric token */ 514 515 check_size_token(3); /* things like "<<=" */ 516 *token.e++ = inbuf_next(); /* if it is only a one-character token, it is 517 * moved here */ 518 *token.e = '\0'; 519 520 switch (*token.s) { 521 case '\n': 522 unary_delim = state->last_u_d; 523 state->last_nl = true; /* remember that we just had a newline */ 524 /* if data has been exhausted, the newline is a dummy. */ 525 ttype = had_eof ? end_of_file : newline; 526 break; 527 528 case '\'': 529 case '"': 530 lex_char_or_string(); 531 ttype = ident; 532 break; 533 534 case '(': 535 case '[': 536 unary_delim = true; 537 ttype = lparen_or_lbracket; 538 break; 539 540 case ')': 541 case ']': 542 ttype = rparen_or_rbracket; 543 break; 544 545 case '#': 546 unary_delim = state->last_u_d; 547 ttype = preprocessing; 548 break; 549 550 case '?': 551 unary_delim = true; 552 ttype = question; 553 break; 554 555 case ':': 556 ttype = colon; 557 unary_delim = true; 558 break; 559 560 case ';': 561 unary_delim = true; 562 ttype = semicolon; 563 break; 564 565 case '{': 566 unary_delim = true; 567 ttype = lbrace; 568 break; 569 570 case '}': 571 unary_delim = true; 572 ttype = rbrace; 573 break; 574 575 case '\f': 576 unary_delim = state->last_u_d; 577 state->last_nl = true; /* remember this, so we can set 'state->col_1' 578 * right */ 579 ttype = form_feed; 580 break; 581 582 case ',': 583 unary_delim = true; 584 ttype = comma; 585 break; 586 587 case '.': 588 unary_delim = false; 589 ttype = period; 590 break; 591 592 case '-': 593 case '+': /* check for -, +, --, ++ */ 594 ttype = state->last_u_d ? unary_op : binary_op; 595 unary_delim = true; 596 597 if (*inp.s == token.s[0]) { 598 /* check for doubled character */ 599 *token.e++ = *inp.s++; 600 /* buffer overflow will be checked at end of loop */ 601 if (state->last_token == ident || 602 state->last_token == rparen_or_rbracket) { 603 ttype = state->last_u_d ? unary_op : postfix_op; 604 /* check for following ++ or -- */ 605 unary_delim = false; 606 } 607 608 } else if (*inp.s == '=') { 609 /* check for operator += */ 610 *token.e++ = *inp.s++; 611 612 } else if (*inp.s == '>') { 613 /* check for operator -> */ 614 *token.e++ = *inp.s++; 615 unary_delim = false; 616 ttype = unary_op; 617 state->want_blank = false; 618 } 619 break; /* buffer overflow will be checked at end of 620 * switch */ 621 622 case '=': 623 if (state->init_or_struct) 624 state->block_init = true; 625 if (*inp.s == '=') { /* == */ 626 *token.e++ = *inp.s++; 627 *token.e = '\0'; 628 } 629 ttype = binary_op; 630 unary_delim = true; 631 break; 632 633 case '>': 634 case '<': 635 case '!': /* ops like <, <<, <=, !=, etc */ 636 if (*inp.s == '>' || *inp.s == '<' || *inp.s == '=') 637 *token.e++ = inbuf_next(); 638 if (*inp.s == '=') 639 *token.e++ = *inp.s++; 640 ttype = state->last_u_d ? unary_op : binary_op; 641 unary_delim = true; 642 break; 643 644 case '*': 645 unary_delim = true; 646 if (!state->last_u_d) { 647 if (*inp.s == '=') 648 *token.e++ = *inp.s++; 649 ttype = binary_op; 650 break; 651 } 652 653 while (*inp.s == '*' || isspace((unsigned char)*inp.s)) { 654 if (*inp.s == '*') 655 token_add_char('*'); 656 inbuf_skip(); 657 } 658 659 if (ps.in_decl) { 660 char *tp = inp.s; 661 662 while (isalpha((unsigned char)*tp) || 663 isspace((unsigned char)*tp)) { 664 if (++tp >= inp.e) 665 inbuf_read_line(); 666 } 667 if (*tp == '(') 668 ps.procname[0] = ' '; 669 } 670 671 ttype = unary_op; 672 break; 673 674 default: 675 if (token.s[0] == '/' && (*inp.s == '*' || *inp.s == '/')) { 676 /* it is start of comment */ 677 *token.e++ = inbuf_next(); 678 679 ttype = comment; 680 unary_delim = state->last_u_d; 681 break; 682 } 683 684 while (token.e[-1] == *inp.s || *inp.s == '=') { 685 /* handle '||', '&&', etc., and also things as in 'int *****i' */ 686 token_add_char(inbuf_next()); 687 } 688 689 ttype = state->last_u_d ? unary_op : binary_op; 690 unary_delim = true; 691 } 692 693 if (inp.s >= inp.e) /* check for input buffer empty */ 694 inbuf_read_line(); 695 696 state->last_u_d = unary_delim; 697 698 check_size_token(1); 699 *token.e = '\0'; 700 701 return lexi_end(ttype); 702 } 703 704 void 705 add_typename(const char *name) 706 { 707 if (typenames.len >= typenames.cap) { 708 typenames.cap = 16 + 2 * typenames.cap; 709 typenames.items = xrealloc(typenames.items, 710 sizeof(typenames.items[0]) * typenames.cap); 711 } 712 713 int pos = bsearch_typenames(name); 714 if (pos >= 0) 715 return; /* already in the list */ 716 717 pos = -(pos + 1); 718 memmove(typenames.items + pos + 1, typenames.items + pos, 719 sizeof(typenames.items[0]) * (typenames.len++ - (unsigned)pos)); 720 typenames.items[pos] = xstrdup(name); 721 } 722
Indexes created Thu Sep 25 16:09:42 GMT 2025