lexi.c revision 1.71
1 /* $NetBSD: lexi.c,v 1.71 2021/10/05 22:09:05 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 1.71 2021/10/05 22:09:05 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 /* INDENT OFF */ 132 static const char num_lex_state[][26] = { 133 /* examples: 134 00 135 s 0xx 136 t 00xaa 137 a 11 101100xxa.. 138 r 11ee0001101lbuuxx.a.pp 139 t.01.e+008bLuxll0Ll.aa.p+0 140 states: ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 141 [0] = "uuiifuufiuuiiuiiiiiuiuuuuu", 142 [1] = "CEIDEHHHIJQ U Q VUVVZZZ", 143 [2] = "DEIDEHHHIJQ U Q VUVVZZZ", 144 [3] = "DEIDEHHHIJ U VUVVZZZ", 145 [4] = "DEJDEHHHJJ U VUVVZZZ", 146 [5] = " U VUVV ", 147 [6] = " K U VUVV ", 148 [7] = " FFF FF U VUVV ", 149 [8] = " f f U VUVV f", 150 [9] = " LLf fL PR Li L f", 151 [10] = " OOf fO S P O i O f", 152 [11] = " FFX ", 153 [12] = " MM M i iiM M ", 154 [13] = " N ", 155 [14] = " G Y ", 156 [15] = "B EE EE T W ", 157 /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 158 }; 159 /* INDENT ON */ 160 161 static const uint8_t num_lex_row[] = { 162 ['0'] = 1, 163 ['1'] = 2, 164 ['2'] = 3, ['3'] = 3, ['4'] = 3, ['5'] = 3, ['6'] = 3, ['7'] = 3, 165 ['8'] = 4, ['9'] = 4, 166 ['A'] = 5, ['a'] = 5, ['C'] = 5, ['c'] = 5, ['D'] = 5, ['d'] = 5, 167 ['B'] = 6, ['b'] = 6, 168 ['E'] = 7, ['e'] = 7, 169 ['F'] = 8, ['f'] = 8, 170 ['L'] = 9, 171 ['l'] = 10, 172 ['P'] = 11, ['p'] = 11, 173 ['U'] = 12, ['u'] = 12, 174 ['X'] = 13, ['x'] = 13, 175 ['+'] = 14, ['-'] = 14, 176 ['.'] = 15, 177 }; 178 179 static char 180 inbuf_peek(void) 181 { 182 return *buf_ptr; 183 } 184 185 void 186 inbuf_skip(void) 187 { 188 buf_ptr++; 189 if (buf_ptr >= buf_end) 190 fill_buffer(); 191 } 192 193 char 194 inbuf_next(void) 195 { 196 char ch = inbuf_peek(); 197 inbuf_skip(); 198 return ch; 199 } 200 201 static void 202 check_size_token(size_t desired_size) 203 { 204 if (token.e + desired_size >= token.l) 205 buf_expand(&token, desired_size); 206 } 207 208 static int 209 cmp_keyword_by_name(const void *key, const void *elem) 210 { 211 return strcmp(key, ((const struct keyword *)elem)->name); 212 } 213 214 static int 215 cmp_type_by_name(const void *key, const void *elem) 216 { 217 return strcmp(key, *((const char *const *)elem)); 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", "rparen", "unary_op", 226 "binary_op", "postfix_op", "question", "case_label", "colon", 227 "semicolon", "lbrace", "rbrace", "ident", "comma", 228 "comment", "switch_expr", "preprocessing", "form_feed", "decl", 229 "keyword_for_if_while", "keyword_do_else", 230 "if_expr", "while_expr", "for_exprs", 231 "stmt", "stmt_list", "keyword_else", "keyword_do", "do_stmt", 232 "if_expr_stmt", "if_expr_stmt_else", "period", "string_prefix", 233 "storage_class", "funcname", "type_def", "keyword_struct_union_enum" 234 }; 235 236 assert(0 <= ttype && ttype < nitems(name)); 237 238 return name[ttype]; 239 } 240 241 static void 242 print_buf(const char *name, const char *s, const char *e) 243 { 244 if (s < e) { 245 debug_printf(" %s ", name); 246 debug_vis_range("\"", s, e, "\""); 247 } 248 } 249 250 static token_type 251 lexi_end(token_type ttype) 252 { 253 debug_printf("in line %d, lexi returns '%s'", 254 line_no, token_type_name(ttype)); 255 print_buf("token", token.s, token.e); 256 print_buf("label", lab.s, lab.e); 257 print_buf("code", code.s, code.e); 258 print_buf("comment", com.s, com.e); 259 debug_printf("\n"); 260 261 return ttype; 262 } 263 #else 264 #define lexi_end(tk) (tk) 265 #endif 266 267 static void 268 lex_number(void) 269 { 270 for (uint8_t s = 'A'; s != 'f' && s != 'i' && s != 'u';) { 271 uint8_t ch = (uint8_t)*buf_ptr; 272 if (ch >= nitems(num_lex_row) || num_lex_row[ch] == 0) 273 break; 274 uint8_t row = num_lex_row[ch]; 275 if (num_lex_state[row][s - 'A'] == ' ') { 276 /*- 277 * num_lex_state[0][s - 'A'] now indicates the type: 278 * f = floating, ch = integer, u = unknown 279 */ 280 break; 281 } 282 s = num_lex_state[row][s - 'A']; 283 check_size_token(1); 284 *token.e++ = inbuf_next(); 285 } 286 } 287 288 static void 289 lex_word(void) 290 { 291 while (isalnum((unsigned char)*buf_ptr) || 292 *buf_ptr == '\\' || 293 *buf_ptr == '_' || *buf_ptr == '$') { 294 /* fill_buffer() terminates buffer with newline */ 295 if (*buf_ptr == '\\') { 296 if (buf_ptr[1] == '\n') { 297 buf_ptr += 2; 298 if (buf_ptr >= buf_end) 299 fill_buffer(); 300 } else 301 break; 302 } 303 check_size_token(1); 304 *token.e++ = inbuf_next(); 305 } 306 } 307 308 static void 309 lex_char_or_string(void) 310 { 311 for (char delim = *token.s;;) { 312 if (*buf_ptr == '\n') { 313 diag(1, "Unterminated literal"); 314 return; 315 } 316 check_size_token(2); 317 *token.e++ = inbuf_next(); 318 if (token.e[-1] == delim) 319 return; 320 if (token.e[-1] == '\\') { 321 if (*buf_ptr == '\n') 322 ++line_no; 323 *token.e++ = inbuf_next(); 324 } 325 } 326 } 327 328 /* 329 * This hack attempts to guess whether the current token is in fact a 330 * declaration keyword -- one that has been defined by typedef. 331 */ 332 static bool 333 probably_typedef(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 (buf_ptr[0] == '*' && buf_ptr[1] != '=') 340 goto maybe; 341 if (isalpha((unsigned char)*buf_ptr)) 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 bool 351 is_typename(void) 352 { 353 if (opt.auto_typedefs) { 354 const char *u; 355 if ((u = strrchr(token.s, '_')) != NULL && strcmp(u, "_t") == 0) 356 return true; 357 } 358 359 if (typenames.len == 0) 360 return false; 361 return bsearch(token.s, typenames.items, (size_t)typenames.len, 362 sizeof(typenames.items[0]), cmp_type_by_name) != NULL; 363 } 364 365 /* Reads the next token, placing it in the global variable "token". */ 366 token_type 367 lexi(struct parser_state *state) 368 { 369 bool unary_delim; /* whether the current token forces a 370 * following operator to be unary */ 371 token_type ttype; 372 373 token.e = token.s; /* point to start of place to save token */ 374 unary_delim = false; 375 state->col_1 = state->last_nl; /* tell world that this token started 376 * in column 1 iff the last thing 377 * scanned was a newline */ 378 state->last_nl = false; 379 380 while (is_hspace(*buf_ptr)) { 381 state->col_1 = false; 382 inbuf_skip(); 383 } 384 385 /* Scan an alphanumeric token */ 386 if (isalnum((unsigned char)*buf_ptr) || 387 *buf_ptr == '_' || *buf_ptr == '$' || 388 (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) { 389 struct keyword *kw; 390 391 if (isdigit((unsigned char)*buf_ptr) || 392 (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) { 393 lex_number(); 394 } else { 395 lex_word(); 396 } 397 *token.e = '\0'; 398 399 if (token.s[0] == 'L' && token.s[1] == '\0' && 400 (*buf_ptr == '"' || *buf_ptr == '\'')) 401 return lexi_end(string_prefix); 402 403 while (is_hspace(inbuf_peek())) 404 inbuf_skip(); 405 state->keyword = kw_0; 406 407 if (state->last_token == keyword_struct_union_enum && 408 state->p_l_follow == 0) { 409 state->last_u_d = true; 410 return lexi_end(decl); 411 } 412 /* 413 * Operator after identifier is binary unless last token was 'struct' 414 */ 415 state->last_u_d = (state->last_token == keyword_struct_union_enum); 416 417 kw = bsearch(token.s, keywords, nitems(keywords), 418 sizeof(keywords[0]), cmp_keyword_by_name); 419 if (kw == NULL) { 420 if (is_typename()) { 421 state->keyword = kw_type; 422 state->last_u_d = true; 423 goto found_typename; 424 } 425 } else { /* we have a keyword */ 426 state->keyword = kw->kind; 427 state->last_u_d = true; 428 switch (kw->kind) { 429 case kw_switch: 430 return lexi_end(switch_expr); 431 case kw_case_or_default: 432 return lexi_end(case_label); 433 case kw_struct_or_union_or_enum: 434 case kw_type: 435 found_typename: 436 if (state->p_l_follow != 0) { 437 /* inside parens: cast, param list, offsetof or sizeof */ 438 state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask; 439 } 440 if (state->last_token == period || state->last_token == unary_op) { 441 state->keyword = kw_0; 442 break; 443 } 444 if (kw != NULL && kw->kind == kw_struct_or_union_or_enum) 445 return lexi_end(keyword_struct_union_enum); 446 if (state->p_l_follow != 0) 447 break; 448 return lexi_end(decl); 449 450 case kw_for_or_if_or_while: 451 return lexi_end(keyword_for_if_while); 452 453 case kw_do_or_else: 454 return lexi_end(keyword_do_else); 455 456 case kw_storage_class: 457 return lexi_end(storage_class); 458 459 case kw_typedef: 460 return lexi_end(type_def); 461 462 default: /* all others are treated like any other 463 * identifier */ 464 return lexi_end(ident); 465 } /* end of switch */ 466 } /* end of if (found_it) */ 467 if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 && 468 !state->in_parameter_declaration && !state->block_init) { 469 char *tp = buf_ptr; 470 while (tp < buf_end) 471 if (*tp++ == ')' && (*tp == ';' || *tp == ',')) 472 goto not_proc; 473 strncpy(state->procname, token.s, sizeof state->procname - 1); 474 if (state->in_decl) 475 state->in_parameter_declaration = true; 476 return lexi_end(funcname); 477 not_proc:; 478 } else if (probably_typedef(state)) { 479 state->keyword = kw_type; 480 state->last_u_d = true; 481 return lexi_end(decl); 482 } 483 if (state->last_token == decl) /* if this is a declared variable, 484 * then following sign is unary */ 485 state->last_u_d = true; /* will make "int a -1" work */ 486 return lexi_end(ident); /* the ident is not in the list */ 487 } /* end of procesing for alpanum character */ 488 489 /* Scan a non-alphanumeric token */ 490 491 check_size_token(3); /* things like "<<=" */ 492 *token.e++ = inbuf_next(); /* if it is only a one-character token, it is 493 * moved here */ 494 *token.e = '\0'; 495 496 switch (*token.s) { 497 case '\n': 498 unary_delim = state->last_u_d; 499 state->last_nl = true; /* remember that we just had a newline */ 500 /* if data has been exhausted, the newline is a dummy. */ 501 ttype = had_eof ? end_of_file : newline; 502 break; 503 504 case '\'': 505 case '"': 506 lex_char_or_string(); 507 ttype = ident; 508 break; 509 510 case '(': 511 case '[': 512 unary_delim = true; 513 ttype = lparen; 514 break; 515 516 case ')': 517 case ']': 518 ttype = rparen; 519 break; 520 521 case '#': 522 unary_delim = state->last_u_d; 523 ttype = preprocessing; 524 break; 525 526 case '?': 527 unary_delim = true; 528 ttype = question; 529 break; 530 531 case ':': 532 ttype = colon; 533 unary_delim = true; 534 break; 535 536 case ';': 537 unary_delim = true; 538 ttype = semicolon; 539 break; 540 541 case '{': 542 unary_delim = true; 543 ttype = lbrace; 544 break; 545 546 case '}': 547 unary_delim = true; 548 ttype = rbrace; 549 break; 550 551 case '\f': 552 unary_delim = state->last_u_d; 553 state->last_nl = true; /* remember this so we can set 'state->col_1' 554 * right */ 555 ttype = form_feed; 556 break; 557 558 case ',': 559 unary_delim = true; 560 ttype = comma; 561 break; 562 563 case '.': 564 unary_delim = false; 565 ttype = period; 566 break; 567 568 case '-': 569 case '+': /* check for -, +, --, ++ */ 570 ttype = state->last_u_d ? unary_op : binary_op; 571 unary_delim = true; 572 573 if (*buf_ptr == token.s[0]) { 574 /* check for doubled character */ 575 *token.e++ = *buf_ptr++; 576 /* buffer overflow will be checked at end of loop */ 577 if (state->last_token == ident || state->last_token == rparen) { 578 ttype = state->last_u_d ? unary_op : postfix_op; 579 /* check for following ++ or -- */ 580 unary_delim = false; 581 } 582 } else if (*buf_ptr == '=') 583 /* check for operator += */ 584 *token.e++ = *buf_ptr++; 585 else if (*buf_ptr == '>') { 586 /* check for operator -> */ 587 *token.e++ = *buf_ptr++; 588 unary_delim = false; 589 ttype = unary_op; 590 state->want_blank = false; 591 } 592 break; /* buffer overflow will be checked at end of 593 * switch */ 594 595 case '=': 596 if (state->in_or_st) 597 state->block_init = true; 598 if (*buf_ptr == '=') { /* == */ 599 *token.e++ = '='; /* Flip =+ to += */ 600 buf_ptr++; 601 *token.e = '\0'; 602 } 603 ttype = binary_op; 604 unary_delim = true; 605 break; 606 /* can drop thru!!! */ 607 608 case '>': 609 case '<': 610 case '!': /* ops like <, <<, <=, !=, etc */ 611 if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') 612 *token.e++ = inbuf_next(); 613 if (*buf_ptr == '=') 614 *token.e++ = *buf_ptr++; 615 ttype = state->last_u_d ? unary_op : binary_op; 616 unary_delim = true; 617 break; 618 619 case '*': 620 unary_delim = true; 621 if (!state->last_u_d) { 622 if (*buf_ptr == '=') 623 *token.e++ = *buf_ptr++; 624 ttype = binary_op; 625 break; 626 } 627 while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) { 628 if (*buf_ptr == '*') { 629 check_size_token(1); 630 *token.e++ = *buf_ptr; 631 } 632 inbuf_skip(); 633 } 634 if (ps.in_decl) { 635 char *tp = buf_ptr; 636 637 while (isalpha((unsigned char)*tp) || 638 isspace((unsigned char)*tp)) { 639 if (++tp >= buf_end) 640 fill_buffer(); 641 } 642 if (*tp == '(') 643 ps.procname[0] = ' '; 644 } 645 ttype = unary_op; 646 break; 647 648 default: 649 if (token.s[0] == '/' && (*buf_ptr == '*' || *buf_ptr == '/')) { 650 /* it is start of comment */ 651 *token.e++ = inbuf_next(); 652 653 ttype = comment; 654 unary_delim = state->last_u_d; 655 break; 656 } 657 while (token.e[-1] == *buf_ptr || *buf_ptr == '=') { 658 /* 659 * handle ||, &&, etc, and also things as in int *****i 660 */ 661 check_size_token(1); 662 *token.e++ = inbuf_next(); 663 } 664 ttype = state->last_u_d ? unary_op : binary_op; 665 unary_delim = true; 666 } 667 668 if (buf_ptr >= buf_end) /* check for input buffer empty */ 669 fill_buffer(); 670 state->last_u_d = unary_delim; 671 check_size_token(1); 672 *token.e = '\0'; 673 return lexi_end(ttype); 674 } 675 676 static int 677 insert_pos(const char *key, const char **arr, unsigned int len) 678 { 679 int lo = 0; 680 int hi = (int)len - 1; 681 682 while (lo <= hi) { 683 int mid = (int)((unsigned)(lo + hi) >> 1); 684 int cmp = strcmp(arr[mid], key); 685 if (cmp < 0) 686 lo = mid + 1; 687 else if (cmp > 0) 688 hi = mid - 1; 689 else 690 return mid; 691 } 692 return -(lo + 1); 693 } 694 695 void 696 add_typename(const char *name) 697 { 698 if (typenames.len >= typenames.cap) { 699 typenames.cap = 16 + 2 * typenames.cap; 700 typenames.items = xrealloc(typenames.items, 701 sizeof(typenames.items[0]) * typenames.cap); 702 } 703 704 int pos = insert_pos(name, typenames.items, typenames.len); 705 if (pos >= 0) 706 return; /* already in the list */ 707 pos = -(pos + 1); 708 memmove(typenames.items + pos + 1, typenames.items + pos, 709 sizeof(typenames.items[0]) * (typenames.len++ - pos)); 710 typenames.items[pos] = xstrdup(name); 711 } 712
Indexes created Tue Sep 30 07:10:03 GMT 2025