lexi.c revision 1.65
1 /* $NetBSD: lexi.c,v 1.65 2021/10/03 20:35:59 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 1.65 2021/10/03 20:35:59 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 {"global", kw_type}, 85 {"goto", kw_jump}, 86 {"if", kw_for_or_if_or_while}, 87 {"imaginary", kw_type}, 88 {"inline", kw_inline_or_restrict}, 89 {"int", kw_type}, 90 {"long", kw_type}, 91 {"offsetof", kw_offsetof}, 92 {"register", kw_storage_class}, 93 {"restrict", kw_inline_or_restrict}, 94 {"return", kw_jump}, 95 {"short", kw_type}, 96 {"signed", kw_type}, 97 {"sizeof", kw_sizeof}, 98 {"static", kw_storage_class}, 99 {"struct", kw_struct_or_union_or_enum}, 100 {"switch", kw_switch}, 101 {"typedef", kw_typedef}, 102 {"union", kw_struct_or_union_or_enum}, 103 {"unsigned", kw_type}, 104 {"void", kw_type}, 105 {"volatile", kw_type}, 106 {"while", kw_for_or_if_or_while} 107 }; 108 109 struct { 110 const char **items; 111 unsigned int len; 112 unsigned int cap; 113 } typenames; 114 115 /* 116 * The transition table below was rewritten by hand from lx's output, given 117 * the following definitions. lx is Katherine Flavel's lexer generator. 118 * 119 * O = /[0-7]/; D = /[0-9]/; NZ = /[1-9]/; 120 * H = /[a-f0-9]/i; B = /[0-1]/; HP = /0x/i; 121 * BP = /0b/i; E = /e[+\-]?/i D+; P = /p[+\-]?/i D+; 122 * FS = /[fl]/i; IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?; 123 * 124 * D+ E FS? -> $float; 125 * D* "." D+ E? FS? -> $float; 126 * D+ "." E? FS? -> $float; HP H+ IS? -> $int; 127 * HP H+ P FS? -> $float; NZ D* IS? -> $int; 128 * HP H* "." H+ P FS? -> $float; "0" O* IS? -> $int; 129 * HP H+ "." P FS -> $float; BP B+ IS? -> $int; 130 */ 131 static const char num_lex_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", 141 [1] = "CEIDEHHHIJQ U Q VUVVZZZ", 142 [2] = "DEIDEHHHIJQ U Q VUVVZZZ", 143 [3] = "DEIDEHHHIJ U VUVVZZZ", 144 [4] = "DEJDEHHHJJ U VUVVZZZ", 145 [5] = " U VUVV ", 146 [6] = " K U VUVV ", 147 [7] = " FFF FF U VUVV ", 148 [8] = " f f U VUVV f", 149 [9] = " LLf fL PR Li L f", 150 [10] = " OOf fO S P O i O f", 151 [11] = " FFX ", 152 [12] = " MM M i iiM M ", 153 [13] = " N ", 154 [14] = " G Y ", 155 [15] = "B EE EE T W ", 156 /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 157 }; 158 159 static const uint8_t num_lex_row[] = { 160 ['0'] = 1, 161 ['1'] = 2, 162 ['2'] = 3, ['3'] = 3, ['4'] = 3, ['5'] = 3, ['6'] = 3, ['7'] = 3, 163 ['8'] = 4, ['9'] = 4, 164 ['A'] = 5, ['a'] = 5, ['C'] = 5, ['c'] = 5, ['D'] = 5, ['d'] = 5, 165 ['B'] = 6, ['b'] = 6, 166 ['E'] = 7, ['e'] = 7, 167 ['F'] = 8, ['f'] = 8, 168 ['L'] = 9, 169 ['l'] = 10, 170 ['P'] = 11, ['p'] = 11, 171 ['U'] = 12, ['u'] = 12, 172 ['X'] = 13, ['x'] = 13, 173 ['+'] = 14, ['-'] = 14, 174 ['.'] = 15, 175 }; 176 177 static char 178 inbuf_peek(void) 179 { 180 return *buf_ptr; 181 } 182 183 static void 184 inbuf_skip(void) 185 { 186 buf_ptr++; 187 if (buf_ptr >= buf_end) 188 fill_buffer(); 189 } 190 191 static char 192 inbuf_next(void) 193 { 194 char ch = inbuf_peek(); 195 inbuf_skip(); 196 return ch; 197 } 198 199 static void 200 check_size_token(size_t desired_size) 201 { 202 if (token.e + desired_size >= token.l) 203 buf_expand(&token, desired_size); 204 } 205 206 static int 207 cmp_keyword_by_name(const void *key, const void *elem) 208 { 209 return strcmp(key, ((const struct keyword *)elem)->name); 210 } 211 212 static int 213 cmp_type_by_name(const void *key, const void *elem) 214 { 215 return strcmp(key, *((const char *const *)elem)); 216 } 217 218 #ifdef debug 219 const char * 220 token_type_name(token_type ttype) 221 { 222 static const char *const name[] = { 223 "end_of_file", "newline", "lparen", "rparen", "unary_op", 224 "binary_op", "postfix_op", "question", "case_label", "colon", 225 "semicolon", "lbrace", "rbrace", "ident", "comma", 226 "comment", "switch_expr", "preprocessing", "form_feed", "decl", 227 "keyword_for_if_while", "keyword_do_else", 228 "if_expr", "while_expr", "for_exprs", 229 "stmt", "stmt_list", "keyword_else", "keyword_do", "do_stmt", 230 "if_expr_stmt", "if_expr_stmt_else", "period", "string_prefix", 231 "storage_class", "funcname", "type_def", "keyword_struct_union_enum" 232 }; 233 234 assert(0 <= ttype && ttype < nitems(name)); 235 236 return name[ttype]; 237 } 238 239 static void 240 print_buf(const char *name, const char *s, const char *e) 241 { 242 if (s < e) { 243 debug_printf(" %s ", name); 244 debug_vis_range("\"", s, e, "\""); 245 } 246 } 247 248 static token_type 249 lexi_end(token_type ttype) 250 { 251 debug_printf("in line %d, lexi returns '%s'", 252 line_no, token_type_name(ttype)); 253 print_buf("token", token.s, token.e); 254 print_buf("label", lab.s, lab.e); 255 print_buf("code", code.s, code.e); 256 print_buf("comment", com.s, com.e); 257 debug_printf("\n"); 258 259 return ttype; 260 } 261 #else 262 # define lexi_end(tk) (tk) 263 #endif 264 265 static void 266 lex_number(void) 267 { 268 for (uint8_t s = 'A'; s != 'f' && s != 'i' && s != 'u'; ) { 269 uint8_t ch = (uint8_t)*buf_ptr; 270 if (ch >= nitems(num_lex_row) || num_lex_row[ch] == 0) 271 break; 272 uint8_t row = num_lex_row[ch]; 273 if (num_lex_state[row][s - 'A'] == ' ') { 274 /* 275 * num_lex_state[0][s - 'A'] now indicates the type: 276 * f = floating, ch = integer, u = unknown 277 */ 278 break; 279 } 280 s = num_lex_state[row][s - 'A']; 281 check_size_token(1); 282 *token.e++ = inbuf_next(); 283 } 284 } 285 286 static void 287 lex_word(void) 288 { 289 while (isalnum((unsigned char)*buf_ptr) || 290 *buf_ptr == '\\' || 291 *buf_ptr == '_' || *buf_ptr == '$') { 292 /* fill_buffer() terminates buffer with newline */ 293 if (*buf_ptr == '\\') { 294 if (buf_ptr[1] == '\n') { 295 buf_ptr += 2; 296 if (buf_ptr >= buf_end) 297 fill_buffer(); 298 } else 299 break; 300 } 301 check_size_token(1); 302 *token.e++ = inbuf_next(); 303 } 304 } 305 306 static void 307 lex_char_or_string(void) 308 { 309 for (char delim = *token.s;;) { 310 if (*buf_ptr == '\n') { 311 diag(1, "Unterminated literal"); 312 return; 313 } 314 check_size_token(2); 315 *token.e++ = inbuf_next(); 316 if (token.e[-1] == delim) 317 return; 318 if (token.e[-1] == '\\') { 319 if (*buf_ptr == '\n') 320 ++line_no; 321 *token.e++ = inbuf_next(); 322 } 323 } 324 } 325 326 /* 327 * This hack attempts to guess whether the current token is in fact a 328 * declaration keyword -- one that has been defined by typedef. 329 */ 330 static bool 331 probably_typedef(const struct parser_state *state) 332 { 333 return state->p_l_follow == 0 && !state->block_init && !state->in_stmt && 334 ((*buf_ptr == '*' && buf_ptr[1] != '=') || 335 isalpha((unsigned char)*buf_ptr)) && 336 (state->last_token == semicolon || state->last_token == lbrace || 337 state->last_token == rbrace); 338 } 339 340 static bool 341 is_typename(void) 342 { 343 if (opt.auto_typedefs) { 344 const char *u; 345 if ((u = strrchr(token.s, '_')) != NULL && strcmp(u, "_t") == 0) 346 return true; 347 } 348 349 if (typenames.len == 0) 350 return false; 351 return bsearch(token.s, typenames.items, (size_t)typenames.len, 352 sizeof(typenames.items[0]), cmp_type_by_name) != NULL; 353 } 354 355 /* Reads the next token, placing it in the global variable "token". */ 356 token_type 357 lexi(struct parser_state *state) 358 { 359 bool unary_delim; /* whether the current token forces a 360 * following operator to be unary */ 361 token_type ttype; 362 363 token.e = token.s; /* point to start of place to save token */ 364 unary_delim = false; 365 state->col_1 = state->last_nl; /* tell world that this token started 366 * in column 1 iff the last thing 367 * scanned was a newline */ 368 state->last_nl = false; 369 370 while (*buf_ptr == ' ' || *buf_ptr == '\t') { /* get rid of blanks */ 371 state->col_1 = false; /* leading blanks imply token is not in column 372 * 1 */ 373 inbuf_skip(); 374 } 375 376 /* Scan an alphanumeric token */ 377 if (isalnum((unsigned char)*buf_ptr) || 378 *buf_ptr == '_' || *buf_ptr == '$' || 379 (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) { 380 struct keyword *kw; 381 382 if (isdigit((unsigned char)*buf_ptr) || 383 (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) { 384 lex_number(); 385 } else { 386 lex_word(); 387 } 388 *token.e = '\0'; 389 390 if (token.s[0] == 'L' && token.s[1] == '\0' && 391 (*buf_ptr == '"' || *buf_ptr == '\'')) 392 return lexi_end(string_prefix); 393 394 while (*buf_ptr == ' ' || *buf_ptr == '\t') /* get rid of blanks */ 395 inbuf_next(); 396 state->keyword = kw_0; 397 398 if (state->last_token == keyword_struct_union_enum && 399 state->p_l_follow == 0) { 400 state->last_u_d = true; 401 return lexi_end(decl); 402 } 403 /* 404 * Operator after identifier is binary unless last token was 'struct' 405 */ 406 state->last_u_d = (state->last_token == keyword_struct_union_enum); 407 408 kw = bsearch(token.s, keywords, nitems(keywords), 409 sizeof(keywords[0]), cmp_keyword_by_name); 410 if (kw == NULL) { 411 if (is_typename()) { 412 state->keyword = kw_type; 413 state->last_u_d = true; 414 goto found_typename; 415 } 416 } else { /* we have a keyword */ 417 state->keyword = kw->kind; 418 state->last_u_d = true; 419 switch (kw->kind) { 420 case kw_switch: 421 return lexi_end(switch_expr); 422 case kw_case_or_default: 423 return lexi_end(case_label); 424 case kw_struct_or_union_or_enum: 425 case kw_type: 426 found_typename: 427 if (state->p_l_follow != 0) { 428 /* inside parens: cast, param list, offsetof or sizeof */ 429 state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask; 430 } 431 if (state->last_token == period || state->last_token == unary_op) { 432 state->keyword = kw_0; 433 break; 434 } 435 if (kw != NULL && kw->kind == kw_struct_or_union_or_enum) 436 return lexi_end(keyword_struct_union_enum); 437 if (state->p_l_follow != 0) 438 break; 439 return lexi_end(decl); 440 441 case kw_for_or_if_or_while: 442 return lexi_end(keyword_for_if_while); 443 444 case kw_do_or_else: 445 return lexi_end(keyword_do_else); 446 447 case kw_storage_class: 448 return lexi_end(storage_class); 449 450 case kw_typedef: 451 return lexi_end(type_def); 452 453 default: /* all others are treated like any other 454 * identifier */ 455 return lexi_end(ident); 456 } /* end of switch */ 457 } /* end of if (found_it) */ 458 if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 && 459 !state->in_parameter_declaration && !state->block_init) { 460 char *tp = buf_ptr; 461 while (tp < buf_end) 462 if (*tp++ == ')' && (*tp == ';' || *tp == ',')) 463 goto not_proc; 464 strncpy(state->procname, token.s, sizeof state->procname - 1); 465 if (state->in_decl) 466 state->in_parameter_declaration = true; 467 return lexi_end(funcname); 468 not_proc:; 469 } else if (probably_typedef(state)) { 470 state->keyword = kw_type; 471 state->last_u_d = true; 472 return lexi_end(decl); 473 } 474 if (state->last_token == decl) /* if this is a declared variable, 475 * then following sign is unary */ 476 state->last_u_d = true; /* will make "int a -1" work */ 477 return lexi_end(ident); /* the ident is not in the list */ 478 } /* end of procesing for alpanum character */ 479 480 /* Scan a non-alphanumeric token */ 481 482 check_size_token(3); /* things like "<<=" */ 483 *token.e++ = inbuf_next(); /* if it is only a one-character token, it is 484 * moved here */ 485 *token.e = '\0'; 486 487 switch (*token.s) { 488 case '\n': 489 unary_delim = state->last_u_d; 490 state->last_nl = true; /* remember that we just had a newline */ 491 /* if data has been exhausted, the newline is a dummy. */ 492 ttype = had_eof ? end_of_file : newline; 493 break; 494 495 case '\'': 496 case '"': 497 lex_char_or_string(); 498 ttype = ident; 499 break; 500 501 case '(': 502 case '[': 503 unary_delim = true; 504 ttype = lparen; 505 break; 506 507 case ')': 508 case ']': 509 ttype = rparen; 510 break; 511 512 case '#': 513 unary_delim = state->last_u_d; 514 ttype = preprocessing; 515 break; 516 517 case '?': 518 unary_delim = true; 519 ttype = question; 520 break; 521 522 case ':': 523 ttype = colon; 524 unary_delim = true; 525 break; 526 527 case ';': 528 unary_delim = true; 529 ttype = semicolon; 530 break; 531 532 case '{': 533 unary_delim = true; 534 ttype = lbrace; 535 break; 536 537 case '}': 538 unary_delim = true; 539 ttype = rbrace; 540 break; 541 542 case 014: /* a form feed */ 543 unary_delim = state->last_u_d; 544 state->last_nl = true; /* remember this so we can set 'state->col_1' 545 * right */ 546 ttype = form_feed; 547 break; 548 549 case ',': 550 unary_delim = true; 551 ttype = comma; 552 break; 553 554 case '.': 555 unary_delim = false; 556 ttype = period; 557 break; 558 559 case '-': 560 case '+': /* check for -, +, --, ++ */ 561 ttype = state->last_u_d ? unary_op : binary_op; 562 unary_delim = true; 563 564 if (*buf_ptr == token.s[0]) { 565 /* check for doubled character */ 566 *token.e++ = *buf_ptr++; 567 /* buffer overflow will be checked at end of loop */ 568 if (state->last_token == ident || state->last_token == rparen) { 569 ttype = state->last_u_d ? unary_op : postfix_op; 570 /* check for following ++ or -- */ 571 unary_delim = false; 572 } 573 } else if (*buf_ptr == '=') 574 /* check for operator += */ 575 *token.e++ = *buf_ptr++; 576 else if (*buf_ptr == '>') { 577 /* check for operator -> */ 578 *token.e++ = *buf_ptr++; 579 unary_delim = false; 580 ttype = unary_op; 581 state->want_blank = false; 582 } 583 break; /* buffer overflow will be checked at end of 584 * switch */ 585 586 case '=': 587 if (state->in_or_st) 588 state->block_init = true; 589 if (*buf_ptr == '=') { /* == */ 590 *token.e++ = '='; /* Flip =+ to += */ 591 buf_ptr++; 592 *token.e = 0; 593 } 594 ttype = binary_op; 595 unary_delim = true; 596 break; 597 /* can drop thru!!! */ 598 599 case '>': 600 case '<': 601 case '!': /* ops like <, <<, <=, !=, etc */ 602 if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') 603 *token.e++ = inbuf_next(); 604 if (*buf_ptr == '=') 605 *token.e++ = *buf_ptr++; 606 ttype = state->last_u_d ? unary_op : binary_op; 607 unary_delim = true; 608 break; 609 610 case '*': 611 unary_delim = true; 612 if (!state->last_u_d) { 613 if (*buf_ptr == '=') 614 *token.e++ = *buf_ptr++; 615 ttype = binary_op; 616 break; 617 } 618 while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) { 619 if (*buf_ptr == '*') { 620 check_size_token(1); 621 *token.e++ = *buf_ptr; 622 } 623 inbuf_skip(); 624 } 625 if (ps.in_decl) { 626 char *tp = buf_ptr; 627 628 while (isalpha((unsigned char)*tp) || 629 isspace((unsigned char)*tp)) { 630 if (++tp >= buf_end) 631 fill_buffer(); 632 } 633 if (*tp == '(') 634 ps.procname[0] = ' '; 635 } 636 ttype = unary_op; 637 break; 638 639 default: 640 if (token.s[0] == '/' && (*buf_ptr == '*' || *buf_ptr == '/')) { 641 /* it is start of comment */ 642 *token.e++ = inbuf_next(); 643 644 ttype = comment; 645 unary_delim = state->last_u_d; 646 break; 647 } 648 while (token.e[-1] == *buf_ptr || *buf_ptr == '=') { 649 /* 650 * handle ||, &&, etc, and also things as in int *****i 651 */ 652 check_size_token(1); 653 *token.e++ = inbuf_next(); 654 } 655 ttype = state->last_u_d ? unary_op : binary_op; 656 unary_delim = true; 657 } 658 659 if (buf_ptr >= buf_end) /* check for input buffer empty */ 660 fill_buffer(); 661 state->last_u_d = unary_delim; 662 check_size_token(1); 663 *token.e = '\0'; 664 return lexi_end(ttype); 665 } 666 667 static int 668 insert_pos(const char *key, const char **arr, unsigned int len) { 669 int lo = 0; 670 int hi = (int)len - 1; 671 672 while (lo <= hi) { 673 int mid = (int)((unsigned)(lo + hi) >> 1); 674 int cmp = strcmp(arr[mid], key); 675 if (cmp < 0) 676 lo = mid + 1; 677 else if (cmp > 0) 678 hi = mid - 1; 679 else 680 return mid; 681 } 682 return -(lo + 1); 683 } 684 685 void 686 add_typename(const char *name) 687 { 688 if (typenames.len >= typenames.cap) { 689 typenames.cap = 16 + 2 * typenames.cap; 690 typenames.items = xrealloc(typenames.items, 691 sizeof(typenames.items[0]) * typenames.cap); 692 } 693 694 int pos = insert_pos(name, typenames.items, typenames.len); 695 if (pos >= 0) 696 return; /* already in the list */ 697 pos = -(pos + 1); 698 memmove(typenames.items + pos + 1, typenames.items + pos, 699 sizeof(typenames.items[0]) * (typenames.len++ - pos)); 700 typenames.items[pos] = xstrdup(name); 701 } 702
Indexes created Thu Sep 25 16:09:42 GMT 2025