lexi.c revision 1.174
1 /* $NetBSD: lexi.c,v 1.174 2023/05/11 10:51:34 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 1.174 2023/05/11 10:51:34 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 <stdlib.h> 52 #include <string.h> 53 54 #include "indent.h" 55 56 /* 57 * While inside lexi_alnum, this constant just marks a type, independently of 58 * the parentheses level. 59 */ 60 #define lsym_type lsym_type_outside_parentheses 61 62 /* must be sorted alphabetically, is used in binary search */ 63 static const struct keyword { 64 const char *name; 65 lexer_symbol lsym; 66 } keywords[] = { 67 {"_Bool", lsym_type}, 68 {"_Complex", lsym_type}, 69 {"_Imaginary", lsym_type}, 70 {"auto", lsym_storage_class}, 71 {"bool", lsym_type}, 72 {"break", lsym_word}, 73 {"case", lsym_case_label}, 74 {"char", lsym_type}, 75 {"complex", lsym_type}, 76 {"const", lsym_type}, 77 {"continue", lsym_word}, 78 {"default", lsym_case_label}, 79 {"do", lsym_do}, 80 {"double", lsym_type}, 81 {"else", lsym_else}, 82 {"enum", lsym_tag}, 83 {"extern", lsym_storage_class}, 84 {"float", lsym_type}, 85 {"for", lsym_for}, 86 {"goto", lsym_word}, 87 {"if", lsym_if}, 88 {"imaginary", lsym_type}, 89 {"inline", lsym_word}, 90 {"int", lsym_type}, 91 {"long", lsym_type}, 92 {"offsetof", lsym_offsetof}, 93 {"register", lsym_storage_class}, 94 {"restrict", lsym_word}, 95 {"return", lsym_return}, 96 {"short", lsym_type}, 97 {"signed", lsym_type}, 98 {"sizeof", lsym_sizeof}, 99 {"static", lsym_storage_class}, 100 {"struct", lsym_tag}, 101 {"switch", lsym_switch}, 102 {"typedef", lsym_typedef}, 103 {"union", lsym_tag}, 104 {"unsigned", lsym_type}, 105 {"void", lsym_type}, 106 {"volatile", lsym_type}, 107 {"while", lsym_while} 108 }; 109 110 static struct { 111 const char **items; 112 unsigned int len; 113 unsigned int cap; 114 } typenames; 115 116 /* 117 * The transition table below was rewritten by hand from lx's output, given 118 * the following definitions. lx is Katherine Flavel's lexer generator. 119 * 120 * O = /[0-7]/; D = /[0-9]/; NZ = /[1-9]/; 121 * H = /[a-f0-9]/i; B = /[0-1]/; HP = /0x/i; 122 * BP = /0b/i; E = /e[+\-]?/i D+; P = /p[+\-]?/i D+; 123 * FS = /[fl]/i; IS = /u/i /(l|L|ll|LL)/? | /(l|L|ll|LL)/ /u/i?; 124 * 125 * D+ E FS? -> $float; 126 * D* "." D+ E? FS? -> $float; 127 * D+ "." E? FS? -> $float; HP H+ IS? -> $int; 128 * HP H+ P FS? -> $float; NZ D* IS? -> $int; 129 * HP H* "." H+ P FS? -> $float; "0" O* IS? -> $int; 130 * HP H+ "." P FS -> $float; BP B+ IS? -> $int; 131 */ 132 /* INDENT OFF */ 133 static const unsigned char lex_number_state[][26] = { 134 /* examples: 135 00 136 s 0xx 137 t 00xaa 138 a 11 101100xxa.. 139 r 11ee0001101lbuuxx.a.pp 140 t.01.e+008bLuxll0Ll.aa.p+0 141 states: ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 142 [0] = "uuiifuufiuuiiuiiiiiuiuuuuu", /* (other) */ 143 [1] = "CEIDEHHHIJQ U Q VUVVZZZ", /* 0 */ 144 [2] = "DEIDEHHHIJQ U Q VUVVZZZ", /* 1 */ 145 [3] = "DEIDEHHHIJ U VUVVZZZ", /* 2 3 4 5 6 7 */ 146 [4] = "DEJDEHHHJJ U VUVVZZZ", /* 8 9 */ 147 [5] = " U VUVV ", /* A a C c D d */ 148 [6] = " K U VUVV ", /* B b */ 149 [7] = " FFF FF U VUVV ", /* E e */ 150 [8] = " f f U VUVV f", /* F f */ 151 [9] = " LLf fL PR Li L f", /* L */ 152 [10] = " OOf fO S P O i O f", /* l */ 153 [11] = " FFX ", /* P p */ 154 [12] = " MM M i iiM M ", /* U u */ 155 [13] = " N ", /* X x */ 156 [14] = " G Y ", /* + - */ 157 [15] = "B EE EE T W ", /* . */ 158 /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */ 159 }; 160 /* INDENT ON */ 161 162 static const unsigned char lex_number_row[] = { 163 ['0'] = 1, 164 ['1'] = 2, 165 ['2'] = 3, ['3'] = 3, ['4'] = 3, ['5'] = 3, ['6'] = 3, ['7'] = 3, 166 ['8'] = 4, ['9'] = 4, 167 ['A'] = 5, ['a'] = 5, ['C'] = 5, ['c'] = 5, ['D'] = 5, ['d'] = 5, 168 ['B'] = 6, ['b'] = 6, 169 ['E'] = 7, ['e'] = 7, 170 ['F'] = 8, ['f'] = 8, 171 ['L'] = 9, 172 ['l'] = 10, 173 ['P'] = 11, ['p'] = 11, 174 ['U'] = 12, ['u'] = 12, 175 ['X'] = 13, ['x'] = 13, 176 ['+'] = 14, ['-'] = 14, 177 ['.'] = 15, 178 }; 179 180 static void 181 check_size_token(size_t desired_size) 182 { 183 if (token.e + desired_size >= token.l) 184 buf_expand(&token, desired_size); 185 } 186 187 static void 188 token_add_char(char ch) 189 { 190 check_size_token(1); 191 *token.e++ = ch; 192 } 193 194 #ifdef debug 195 static const char * 196 lsym_name(lexer_symbol sym) 197 { 198 static const char *const name[] = { 199 "eof", 200 "preprocessing", 201 "newline", 202 "form_feed", 203 "comment", 204 "lparen_or_lbracket", 205 "rparen_or_rbracket", 206 "lbrace", 207 "rbrace", 208 "period", 209 "unary_op", 210 "binary_op", 211 "postfix_op", 212 "question", 213 "colon", 214 "comma", 215 "semicolon", 216 "typedef", 217 "storage_class", 218 "type_outside_parentheses", 219 "type_in_parentheses", 220 "tag", 221 "case_label", 222 "sizeof", 223 "offsetof", 224 "word", 225 "funcname", 226 "do", 227 "else", 228 "for", 229 "if", 230 "switch", 231 "while", 232 "return", 233 }; 234 235 return name[sym]; 236 } 237 238 static void 239 debug_print_buf(const char *name, const struct buffer *buf) 240 { 241 if (buf->s < buf->e) { 242 debug_printf("%s ", name); 243 debug_vis_range("\"", buf->s, buf->e, "\"\n"); 244 } 245 } 246 247 static bool 248 debug_full_parser_state(void) 249 { 250 return true; 251 } 252 253 #define debug_ps_bool(name) \ 254 if (ps.name != prev_ps.name) \ 255 debug_println("[%c] -> [%c] ps." #name, \ 256 prev_ps.name ? 'x' : ' ', ps.name ? 'x' : ' '); \ 257 else if (debug_full_parser_state()) \ 258 debug_println(" [%c] ps." #name, ps.name ? 'x' : ' ') 259 #define debug_ps_int(name) \ 260 if (ps.name != prev_ps.name) \ 261 debug_println("%3d -> %3d ps." #name, prev_ps.name, ps.name); \ 262 else if (debug_full_parser_state()) \ 263 debug_println(" %3d ps." #name, ps.name) 264 265 static bool 266 ps_paren_has_changed(const struct parser_state *prev_ps) 267 { 268 const paren_level_props *prev = prev_ps->paren, *curr = ps.paren; 269 270 if (prev_ps->nparen != ps.nparen) 271 return true; 272 273 for (int i = 0; i < ps.nparen; i++) { 274 if (curr[i].indent != prev[i].indent || 275 curr[i].maybe_cast != prev[i].maybe_cast || 276 curr[i].no_cast != prev[i].no_cast) 277 return true; 278 } 279 return false; 280 } 281 282 static void 283 debug_ps_paren(const struct parser_state *prev_ps) 284 { 285 if (!debug_full_parser_state() && !ps_paren_has_changed(prev_ps)) 286 return; 287 288 debug_printf(" ps.paren:"); 289 for (int i = 0; i < ps.nparen; i++) { 290 const paren_level_props *props = ps.paren + i; 291 const char *cast = props->no_cast ? "(no cast)" 292 : props->maybe_cast ? "(cast)" 293 : ""; 294 debug_printf(" %s%d", cast, props->indent); 295 } 296 if (ps.nparen == 0) 297 debug_printf(" none"); 298 debug_println(""); 299 } 300 301 static void 302 debug_lexi(lexer_symbol lsym) 303 { 304 /* 305 * Watch out for 'rolled back parser state' in the debug output; the 306 * differences around these are unreliable. 307 */ 308 static struct parser_state prev_ps; 309 310 debug_println(""); 311 debug_printf("line %d: %s", line_no, lsym_name(lsym)); 312 debug_vis_range(" \"", token.s, token.e, "\"\n"); 313 314 debug_print_buf("label", &lab); 315 debug_print_buf("code", &code); 316 debug_print_buf("comment", &com); 317 318 debug_println(" ps.prev_token = %s", lsym_name(ps.prev_token)); 319 debug_ps_bool(next_col_1); 320 debug_ps_bool(curr_col_1); 321 debug_ps_bool(next_unary); 322 debug_ps_bool(is_function_definition); 323 debug_ps_bool(want_blank); 324 debug_ps_bool(force_nl); 325 debug_ps_int(line_start_nparen); 326 debug_ps_int(nparen); 327 debug_ps_paren(&prev_ps); 328 329 debug_ps_int(comment_delta); 330 debug_ps_int(n_comment_delta); 331 debug_ps_int(com_ind); 332 333 debug_ps_bool(block_init); 334 debug_ps_int(block_init_level); 335 debug_ps_bool(init_or_struct); 336 337 debug_ps_int(ind_level); 338 debug_ps_int(ind_level_follow); 339 340 debug_ps_int(decl_level); 341 debug_ps_bool(decl_on_line); 342 debug_ps_bool(in_decl); 343 debug_ps_int(just_saw_decl); 344 debug_ps_bool(in_func_def_params); 345 debug_ps_bool(decl_indent_done); 346 347 debug_ps_bool(in_stmt_or_decl); 348 debug_ps_bool(in_stmt_cont); 349 debug_ps_bool(is_case_label); 350 351 prev_ps = ps; 352 } 353 #endif 354 355 static lexer_symbol 356 lexi_end(lexer_symbol lsym) 357 { 358 #ifdef debug 359 debug_lexi(lsym); 360 #endif 361 return lsym; 362 } 363 364 static void 365 lex_number(void) 366 { 367 for (unsigned char s = 'A'; s != 'f' && s != 'i' && s != 'u';) { 368 unsigned char ch = (unsigned char)inp_peek(); 369 if (ch >= array_length(lex_number_row) || lex_number_row[ch] == 0) 370 break; 371 372 unsigned char row = lex_number_row[ch]; 373 if (lex_number_state[row][s - 'A'] == ' ') { 374 /*- 375 * lex_number_state[0][s - 'A'] now indicates the type: 376 * f = floating, i = integer, u = unknown 377 */ 378 return; 379 } 380 381 s = lex_number_state[row][s - 'A']; 382 token_add_char(inp_next()); 383 } 384 } 385 386 static bool 387 is_identifier_start(char ch) 388 { 389 return ch_isalpha(ch) || ch == '_' || ch == '$'; 390 } 391 392 static bool 393 is_identifier_part(char ch) 394 { 395 return ch_isalnum(ch) || ch == '_' || ch == '$'; 396 } 397 398 static void 399 lex_word(void) 400 { 401 for (;;) { 402 if (is_identifier_part(inp_peek())) 403 token_add_char(inp_next()); 404 else if (inp_peek() == '\\' && inp_lookahead(1) == '\n') { 405 inp_skip(); 406 inp_skip(); 407 } else 408 return; 409 } 410 } 411 412 static void 413 lex_char_or_string(void) 414 { 415 for (char delim = token.e[-1];;) { 416 if (inp_peek() == '\n') { 417 diag(1, "Unterminated literal"); 418 return; 419 } 420 421 token_add_char(inp_next()); 422 if (token.e[-1] == delim) 423 return; 424 425 if (token.e[-1] == '\\') { 426 if (inp_peek() == '\n') 427 ++line_no; 428 token_add_char(inp_next()); 429 } 430 } 431 } 432 433 /* Guess whether the current token is a declared type. */ 434 static bool 435 probably_typename(void) 436 { 437 if (ps.prev_token == lsym_storage_class) 438 return true; 439 if (ps.block_init) 440 return false; 441 if (ps.in_stmt_or_decl) /* XXX: this condition looks incorrect */ 442 return false; 443 if (inp_peek() == '*' && inp_lookahead(1) != '=') 444 goto maybe; 445 /* XXX: is_identifier_start */ 446 if (ch_isalpha(inp_peek())) 447 goto maybe; 448 return false; 449 maybe: 450 return ps.prev_token == lsym_semicolon || 451 ps.prev_token == lsym_lbrace || 452 ps.prev_token == lsym_rbrace; 453 } 454 455 static int 456 bsearch_typenames(const char *key) 457 { 458 const char **arr = typenames.items; 459 int lo = 0; 460 int hi = (int)typenames.len - 1; 461 462 while (lo <= hi) { 463 int mid = (int)((unsigned)(lo + hi) >> 1); 464 int cmp = strcmp(arr[mid], key); 465 if (cmp < 0) 466 lo = mid + 1; 467 else if (cmp > 0) 468 hi = mid - 1; 469 else 470 return mid; 471 } 472 return -(lo + 1); 473 } 474 475 static bool 476 is_typename(void) 477 { 478 if (opt.auto_typedefs && 479 token.e - token.s >= 2 && memcmp(token.e - 2, "_t", 2) == 0) 480 return true; 481 482 return bsearch_typenames(token.s) >= 0; 483 } 484 485 static int 486 cmp_keyword_by_name(const void *key, const void *elem) 487 { 488 return strcmp(key, ((const struct keyword *)elem)->name); 489 } 490 491 /* 492 * Looking at something like 'function_name(...)' in a line, guess whether 493 * this starts a function definition or a declaration. 494 */ 495 static bool 496 probably_looking_at_definition(void) 497 { 498 int paren_level = 0; 499 for (const char *p = inp_p(), *e = inp_line_end(); p < e; p++) { 500 if (*p == '(') 501 paren_level++; 502 if (*p == ')' && --paren_level == 0) { 503 p++; 504 505 while (p < e && (ch_isspace(*p) || is_identifier_part(*p))) 506 p++; /* '__dead' or '__unused' */ 507 508 if (p == e) /* func(...) */ 509 break; 510 if (*p == ';') /* func(...); */ 511 return false; 512 if (*p == ',') /* double abs(), pi; */ 513 return false; 514 if (*p == '(') /* func(...) __attribute__((...)) */ 515 paren_level++; /* func(...) __printflike(...) */ 516 else 517 break; /* func(...) { ... */ 518 } 519 } 520 521 /* 522 * To further reduce the cases where indent wrongly treats an incomplete 523 * function declaration as a function definition, thus adding a newline 524 * before the function name, it may be worth looking for parameter names, 525 * as these are often omitted in function declarations and only included 526 * in function definitions. Or just increase the lookahead to more than 527 * just the current line of input, until the next '{'. 528 */ 529 return true; 530 } 531 532 /* Read an alphanumeric token into 'token', or return lsym_eof. */ 533 static lexer_symbol 534 lexi_alnum(void) 535 { 536 if (ch_isdigit(inp_peek()) || 537 (inp_peek() == '.' && ch_isdigit(inp_lookahead(1)))) { 538 lex_number(); 539 } else if (is_identifier_start(inp_peek())) { 540 lex_word(); 541 542 if (token.s[0] == 'L' && token.e - token.s == 1 && 543 (inp_peek() == '"' || inp_peek() == '\'')) { 544 token_add_char(inp_next()); 545 lex_char_or_string(); 546 ps.next_unary = false; 547 548 check_size_token(1); 549 *token.e = '\0'; 550 551 return lsym_word; 552 } 553 } else 554 return lsym_eof; /* just as a placeholder */ 555 556 *token.e = '\0'; 557 558 while (ch_isblank(inp_peek())) 559 inp_skip(); 560 561 ps.next_unary = ps.prev_token == lsym_tag; /* for 'struct s *' */ 562 563 if (ps.prev_token == lsym_tag && ps.nparen == 0) 564 return lsym_type_outside_parentheses; 565 566 const struct keyword *kw = bsearch(token.s, keywords, 567 array_length(keywords), sizeof(keywords[0]), cmp_keyword_by_name); 568 bool is_type = false; 569 if (kw == NULL) { 570 if (is_typename()) { 571 is_type = true; 572 ps.next_unary = true; 573 if (ps.in_enum == in_enum_enum) 574 ps.in_enum = in_enum_type; 575 goto found_typename; 576 } 577 578 } else { /* we have a keyword */ 579 is_type = kw->lsym == lsym_type; 580 ps.next_unary = true; 581 if (kw->lsym != lsym_tag && kw->lsym != lsym_type) 582 return kw->lsym; 583 584 found_typename: 585 if (ps.nparen > 0) { 586 /* inside parentheses: cast, param list, offsetof or sizeof */ 587 if (!ps.paren[ps.nparen - 1].no_cast) 588 ps.paren[ps.nparen - 1].maybe_cast = true; 589 } 590 if (ps.prev_token != lsym_period && ps.prev_token != lsym_unary_op) { 591 if (kw != NULL && kw->lsym == lsym_tag) { 592 if (token.s[0] == 'e' /* enum */) 593 ps.in_enum = in_enum_enum; 594 return lsym_tag; 595 } 596 if (ps.nparen == 0) 597 return lsym_type_outside_parentheses; 598 } 599 } 600 601 if (inp_peek() == '(' && ps.tos <= 1 && ps.ind_level == 0 && 602 !ps.in_func_def_params && !ps.block_init) { 603 604 if (ps.nparen == 0 && probably_looking_at_definition()) { 605 ps.is_function_definition = true; 606 if (ps.in_decl) 607 ps.in_func_def_params = true; 608 return lsym_funcname; 609 } 610 611 } else if (ps.nparen == 0 && probably_typename()) { 612 ps.next_unary = true; 613 return lsym_type_outside_parentheses; 614 } 615 616 return is_type ? lsym_type_in_parentheses : lsym_word; 617 } 618 619 static bool 620 is_asterisk_unary(void) 621 { 622 if (ps.next_unary || ps.in_func_def_params) 623 return true; 624 if (ps.prev_token == lsym_word || 625 ps.prev_token == lsym_rparen_or_rbracket) 626 return false; 627 return ps.in_decl && ps.nparen > 0; 628 } 629 630 static void 631 lex_asterisk_unary(void) 632 { 633 while (inp_peek() == '*' || ch_isspace(inp_peek())) { 634 if (inp_peek() == '*') 635 token_add_char('*'); 636 inp_skip(); 637 } 638 639 if (ps.in_decl) { 640 const char *tp = inp_p(), *e = inp_line_end(); 641 642 while (tp < e) { 643 if (ch_isspace(*tp)) 644 tp++; 645 else if (is_identifier_start(*tp)) { 646 tp++; 647 while (tp < e && is_identifier_part(*tp)) 648 tp++; 649 } else 650 break; 651 } 652 653 if (tp < e && *tp == '(') 654 ps.is_function_definition = true; 655 } 656 } 657 658 /* Reads the next token, placing it in the global variable "token". */ 659 lexer_symbol 660 lexi(void) 661 { 662 token.e = token.s; 663 ps.curr_col_1 = ps.next_col_1; 664 ps.next_col_1 = false; 665 666 while (ch_isblank(inp_peek())) { 667 ps.curr_col_1 = false; 668 inp_skip(); 669 } 670 671 lexer_symbol alnum_lsym = lexi_alnum(); 672 if (alnum_lsym != lsym_eof) 673 return lexi_end(alnum_lsym); 674 675 /* Scan a non-alphanumeric token */ 676 677 check_size_token(3); /* for things like "<<=" */ 678 *token.e++ = inp_next(); 679 *token.e = '\0'; 680 681 lexer_symbol lsym; 682 bool next_unary; 683 684 switch (token.e[-1]) { 685 686 /* INDENT OFF */ 687 case '(': 688 case '[': lsym = lsym_lparen_or_lbracket; next_unary = true; break; 689 case ')': 690 case ']': lsym = lsym_rparen_or_rbracket; next_unary = false; break; 691 case '?': lsym = lsym_question; next_unary = true; break; 692 case ':': lsym = lsym_colon; next_unary = true; break; 693 case ';': lsym = lsym_semicolon; next_unary = true; break; 694 case '{': lsym = lsym_lbrace; next_unary = true; break; 695 case '}': lsym = lsym_rbrace; next_unary = true; break; 696 case ',': lsym = lsym_comma; next_unary = true; break; 697 case '.': lsym = lsym_period; next_unary = false; break; 698 /* INDENT ON */ 699 700 case '\n': 701 /* if data has been exhausted, the '\n' is a dummy. */ 702 lsym = had_eof ? lsym_eof : lsym_newline; 703 next_unary = ps.next_unary; 704 ps.next_col_1 = true; 705 break; 706 707 case '\f': 708 lsym = lsym_form_feed; 709 next_unary = ps.next_unary; 710 ps.next_col_1 = true; 711 break; 712 713 case '#': 714 lsym = lsym_preprocessing; 715 next_unary = ps.next_unary; 716 break; 717 718 case '\'': 719 case '"': 720 lex_char_or_string(); 721 lsym = lsym_word; 722 next_unary = false; 723 break; 724 725 case '-': 726 case '+': 727 lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op; 728 next_unary = true; 729 730 if (inp_peek() == token.e[-1]) { /* '++' or '--' */ 731 *token.e++ = inp_next(); 732 if (ps.prev_token == lsym_word || 733 ps.prev_token == lsym_rparen_or_rbracket) { 734 lsym = ps.next_unary ? lsym_unary_op : lsym_postfix_op; 735 next_unary = false; 736 } 737 738 } else if (inp_peek() == '=') { /* '+=' or '-=' */ 739 *token.e++ = inp_next(); 740 741 } else if (inp_peek() == '>') { /* '->' */ 742 *token.e++ = inp_next(); 743 lsym = lsym_unary_op; 744 next_unary = false; 745 ps.want_blank = false; 746 } 747 break; 748 749 case '=': 750 if (ps.init_or_struct) 751 ps.block_init = true; 752 if (inp_peek() == '=') { /* == */ 753 *token.e++ = inp_next(); 754 *token.e = '\0'; 755 } 756 lsym = lsym_binary_op; 757 next_unary = true; 758 break; 759 760 case '>': 761 case '<': 762 case '!': /* ops like <, <<, <=, !=, etc */ 763 if (inp_peek() == '>' || inp_peek() == '<' || inp_peek() == '=') 764 *token.e++ = inp_next(); 765 if (inp_peek() == '=') 766 *token.e++ = inp_next(); 767 lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op; 768 next_unary = true; 769 break; 770 771 case '*': 772 if (is_asterisk_unary()) { 773 lex_asterisk_unary(); 774 lsym = lsym_unary_op; 775 next_unary = true; 776 } else { 777 if (inp_peek() == '=') 778 *token.e++ = inp_next(); 779 lsym = lsym_binary_op; 780 next_unary = true; 781 } 782 break; 783 784 default: 785 if (token.e[-1] == '/' && (inp_peek() == '*' || inp_peek() == '/')) { 786 *token.e++ = inp_next(); 787 lsym = lsym_comment; 788 next_unary = ps.next_unary; 789 break; 790 } 791 792 /* handle '||', '&&', etc., and also things as in 'int *****i' */ 793 while (token.e[-1] == inp_peek() || inp_peek() == '=') 794 token_add_char(inp_next()); 795 796 lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op; 797 next_unary = true; 798 } 799 800 if (ps.in_enum == in_enum_enum || ps.in_enum == in_enum_type) 801 ps.in_enum = lsym == lsym_lbrace ? in_enum_brace : in_enum_no; 802 if (lsym == lsym_rbrace) 803 ps.in_enum = in_enum_no; 804 805 ps.next_unary = next_unary; 806 807 check_size_token(1); 808 *token.e = '\0'; 809 810 return lexi_end(lsym); 811 } 812 813 void 814 register_typename(const char *name) 815 { 816 if (typenames.len >= typenames.cap) { 817 typenames.cap = 16 + 2 * typenames.cap; 818 typenames.items = xrealloc(typenames.items, 819 sizeof(typenames.items[0]) * typenames.cap); 820 } 821 822 int pos = bsearch_typenames(name); 823 if (pos >= 0) 824 return; /* already in the list */ 825 826 pos = -(pos + 1); 827 memmove(typenames.items + pos + 1, typenames.items + pos, 828 sizeof(typenames.items[0]) * (typenames.len++ - (unsigned)pos)); 829 typenames.items[pos] = xstrdup(name); 830 } 831
Indexes created Mon Oct 27 08:10:08 GMT 2025