Annotation of src/usr.bin/indent/lexi.c, Revision 1.114
1.114 ! rillig 1: /* $NetBSD: lexi.c,v 1.113 2021/10/29 21:31:29 rillig Exp $ */
1.3 tls 2:
1.16 kamil 3: /*-
4: * SPDX-License-Identifier: BSD-4-Clause
5: *
6: * Copyright (c) 1985 Sun Microsystems, Inc.
1.5 mrg 7: * Copyright (c) 1980, 1993
8: * The Regents of the University of California. All rights reserved.
1.1 cgd 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:
1.16 kamil 40: #if 0
41: static char sccsid[] = "@(#)lexi.c 8.1 (Berkeley) 6/6/93";
42: #endif
43:
1.6 lukem 44: #include <sys/cdefs.h>
1.16 kamil 45: #if defined(__NetBSD__)
1.114 ! rillig 46: __RCSID("$NetBSD: lexi.c,v 1.113 2021/10/29 21:31:29 rillig Exp $");
1.16 kamil 47: #elif defined(__FreeBSD__)
48: __FBSDID("$FreeBSD: head/usr.bin/indent/lexi.c 337862 2018-08-15 18:19:45Z pstef $");
49: #endif
1.1 cgd 50:
1.93 rillig 51: #include <sys/param.h>
1.20 rillig 52: #include <assert.h>
1.1 cgd 53: #include <stdio.h>
54: #include <ctype.h>
55: #include <stdlib.h>
56: #include <string.h>
1.16 kamil 57:
58: #include "indent.h"
1.1 cgd 59:
1.60 rillig 60: /* must be sorted alphabetically, is used in binary search */
1.62 rillig 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},
1.97 rillig 77: {"do", kw_do},
1.62 rillig 78: {"double", kw_type},
1.97 rillig 79: {"else", kw_else},
1.62 rillig 80: {"enum", kw_struct_or_union_or_enum},
81: {"extern", kw_storage_class},
82: {"float", kw_type},
1.98 rillig 83: {"for", kw_for},
1.62 rillig 84: {"goto", kw_jump},
1.98 rillig 85: {"if", kw_if},
1.62 rillig 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},
1.98 rillig 105: {"while", kw_while}
1.1 cgd 106: };
107:
1.84 rillig 108: static struct {
1.64 rillig 109: const char **items;
110: unsigned int len;
111: unsigned int cap;
112: } typenames;
1.16 kamil 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: */
1.71 rillig 130: /* INDENT OFF */
1.82 rillig 131: static const unsigned char lex_number_state[][26] = {
1.16 kamil 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 */
1.83 rillig 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 ", /* . */
1.16 kamil 156: /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */
1.1 cgd 157: };
1.71 rillig 158: /* INDENT ON */
1.1 cgd 159:
1.82 rillig 160: static const uint8_t lex_number_row[] = {
1.56 rillig 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: };
1.36 rillig 177:
1.32 rillig 178: static char
179: inbuf_peek(void)
180: {
1.78 rillig 181: return *inp.s;
1.32 rillig 182: }
183:
1.66 rillig 184: void
1.32 rillig 185: inbuf_skip(void)
186: {
1.78 rillig 187: inp.s++;
188: if (inp.s >= inp.e)
1.81 rillig 189: inbuf_read_line();
1.32 rillig 190: }
191:
1.66 rillig 192: char
1.32 rillig 193: inbuf_next(void)
194: {
195: char ch = inbuf_peek();
196: inbuf_skip();
197: return ch;
198: }
199:
1.25 rillig 200: static void
201: check_size_token(size_t desired_size)
202: {
1.58 rillig 203: if (token.e + desired_size >= token.l)
204: buf_expand(&token, desired_size);
1.25 rillig 205: }
206:
1.87 rillig 207: static void
208: token_add_char(char ch)
209: {
210: check_size_token(1);
211: *token.e++ = ch;
212: }
213:
1.16 kamil 214: static int
1.62 rillig 215: cmp_keyword_by_name(const void *key, const void *elem)
1.16 kamil 216: {
1.62 rillig 217: return strcmp(key, ((const struct keyword *)elem)->name);
1.27 rillig 218: }
219:
1.20 rillig 220: #ifdef debug
1.100 rillig 221: static const char *
222: lsym_name(lexer_symbol sym)
1.20 rillig 223: {
224: static const char *const name[] = {
1.100 rillig 225: "eof",
226: "preprocessing",
227: "newline",
228: "form_feed",
229: "comment",
230: "lparen_or_lbracket",
231: "rparen_or_rbracket",
232: "lbrace",
233: "rbrace",
234: "period",
235: "unary_op",
236: "binary_op",
237: "postfix_op",
238: "question",
239: "colon",
240: "comma",
241: "semicolon",
242: "typedef",
243: "storage_class",
244: "type",
245: "tag",
246: "case_label",
247: "string_prefix",
248: "ident",
249: "funcname",
250: "do",
251: "else",
252: "for",
253: "if",
254: "switch",
255: "while",
1.20 rillig 256: };
257:
1.100 rillig 258: assert(array_length(name) == (int)lsym_while + 1);
1.20 rillig 259:
1.100 rillig 260: return name[sym];
1.20 rillig 261: }
262:
1.101 rillig 263: static const char *
1.103 rillig 264: kw_name(enum keyword_kind kw)
265: {
1.101 rillig 266: static const char *name[] = {
267: "0",
268: "offsetof",
269: "sizeof",
270: "struct_or_union_or_enum",
271: "type",
272: "for",
273: "if",
274: "while",
275: "do",
276: "else",
277: "switch",
278: "case_or_default",
279: "jump",
280: "storage_class",
281: "typedef",
282: "inline_or_restrict",
283: };
284:
285: return name[kw];
286: }
287:
1.20 rillig 288: static void
1.72 rillig 289: debug_print_buf(const char *name, const struct buffer *buf)
1.20 rillig 290: {
1.72 rillig 291: if (buf->s < buf->e) {
1.101 rillig 292: debug_printf("%s ", name);
293: debug_vis_range("\"", buf->s, buf->e, "\"\n");
1.20 rillig 294: }
295: }
296:
1.112 rillig 297: #define debug_ps_bool(name) \
1.113 rillig 298: if (ps.name != prev_ps.name) \
299: debug_println("[%c] ps." #name, ps.name ? 'x' : ' ')
1.112 rillig 300: #define debug_ps_int(name) \
1.113 rillig 301: if (ps.name != prev_ps.name) \
302: debug_println("%3d ps." #name, ps.name)
1.112 rillig 303: #define debug_ps_keyword(name) \
304: if (ps.name != kw_0) \
1.113 rillig 305: debug_println(" ps." #name " = %s", kw_name(ps.name))
1.112 rillig 306:
1.101 rillig 307: static void
1.107 rillig 308: debug_lexi(lexer_symbol lsym)
1.20 rillig 309: {
1.113 rillig 310: /*
311: * Watch out for 'rolled back parser state' in the debug output; the
312: * differences around these are unreliable.
313: */
314: static struct parser_state prev_ps;
315:
1.104 rillig 316: debug_println("");
1.101 rillig 317: debug_printf("line %d\n", line_no);
1.72 rillig 318: debug_print_buf("label", &lab);
319: debug_print_buf("code", &code);
320: debug_print_buf("comment", &com);
1.114 ! rillig 321: debug_printf("lexi: %s", lsym_name(lsym));
1.112 rillig 322: debug_vis_range(" \"", token.s, token.e, "\"\n");
323:
324: // prev_token
325: debug_ps_bool(prev_newline);
326: debug_ps_bool(prev_col_1);
327: debug_ps_keyword(prev_keyword);
328: debug_ps_keyword(curr_keyword);
329: debug_ps_bool(next_unary);
330: // procname
331: debug_ps_bool(want_blank);
332: debug_ps_int(paren_level);
333: debug_ps_int(p_l_follow);
334: // paren_indents
335: debug_ps_int(cast_mask);
336: debug_ps_int(not_cast_mask);
337:
338: debug_ps_int(comment_delta);
339: debug_ps_int(n_comment_delta);
340: debug_ps_int(com_ind);
341:
342: debug_ps_bool(block_init);
343: debug_ps_int(block_init_level);
344: debug_ps_bool(init_or_struct);
345:
346: debug_ps_int(ind_level);
347: debug_ps_int(ind_level_follow);
348:
349: debug_ps_int(decl_nest);
350: debug_ps_bool(decl_on_line);
351: debug_ps_bool(in_decl);
352: debug_ps_int(just_saw_decl);
353: debug_ps_bool(in_parameter_declaration);
354: debug_ps_bool(decl_indent_done);
355:
356: debug_ps_bool(in_stmt);
357: debug_ps_bool(ind_stmt);
358: debug_ps_bool(is_case_label);
359:
360: debug_ps_bool(search_stmt);
1.113 rillig 361:
362: prev_ps = ps;
1.101 rillig 363: }
1.96 rillig 364: #endif
1.20 rillig 365:
1.104 rillig 366: /* ARGSUSED */
1.101 rillig 367: static lexer_symbol
1.107 rillig 368: lexi_end(lexer_symbol lsym)
1.101 rillig 369: {
370: #ifdef debug
1.107 rillig 371: debug_lexi(lsym);
1.101 rillig 372: #endif
1.100 rillig 373: return lsym;
1.20 rillig 374: }
375:
1.43 rillig 376: static void
377: lex_number(void)
378: {
1.71 rillig 379: for (uint8_t s = 'A'; s != 'f' && s != 'i' && s != 'u';) {
1.78 rillig 380: uint8_t ch = (uint8_t)*inp.s;
1.94 rillig 381: if (ch >= array_length(lex_number_row) || lex_number_row[ch] == 0)
1.56 rillig 382: break;
1.75 rillig 383:
1.82 rillig 384: uint8_t row = lex_number_row[ch];
385: if (lex_number_state[row][s - 'A'] == ' ') {
1.71 rillig 386: /*-
1.82 rillig 387: * lex_number_state[0][s - 'A'] now indicates the type:
1.74 rillig 388: * f = floating, i = integer, u = unknown
1.56 rillig 389: */
1.43 rillig 390: break;
391: }
1.75 rillig 392:
1.82 rillig 393: s = lex_number_state[row][s - 'A'];
1.87 rillig 394: token_add_char(inbuf_next());
1.43 rillig 395: }
396: }
397:
398: static void
399: lex_word(void)
400: {
1.78 rillig 401: while (isalnum((unsigned char)*inp.s) ||
1.95 rillig 402: *inp.s == '\\' ||
403: *inp.s == '_' || *inp.s == '$') {
1.75 rillig 404:
1.78 rillig 405: if (*inp.s == '\\') {
406: if (inp.s[1] == '\n') {
407: inp.s += 2;
408: if (inp.s >= inp.e)
1.81 rillig 409: inbuf_read_line();
1.43 rillig 410: } else
411: break;
412: }
1.75 rillig 413:
1.87 rillig 414: token_add_char(inbuf_next());
1.43 rillig 415: }
416: }
417:
418: static void
419: lex_char_or_string(void)
420: {
1.52 rillig 421: for (char delim = *token.s;;) {
1.78 rillig 422: if (*inp.s == '\n') {
1.52 rillig 423: diag(1, "Unterminated literal");
424: return;
425: }
1.75 rillig 426:
1.87 rillig 427: token_add_char(inbuf_next());
1.52 rillig 428: if (token.e[-1] == delim)
429: return;
1.75 rillig 430:
1.52 rillig 431: if (token.e[-1] == '\\') {
1.78 rillig 432: if (*inp.s == '\n')
1.52 rillig 433: ++line_no;
1.87 rillig 434: token_add_char(inbuf_next());
1.52 rillig 435: }
436: }
1.43 rillig 437: }
438:
1.84 rillig 439: /* Guess whether the current token is a declared type. */
1.57 rillig 440: static bool
1.107 rillig 441: probably_typename(void)
1.57 rillig 442: {
1.109 rillig 443: if (ps.p_l_follow > 0)
1.70 rillig 444: return false;
1.107 rillig 445: if (ps.block_init || ps.in_stmt)
1.70 rillig 446: return false;
1.78 rillig 447: if (inp.s[0] == '*' && inp.s[1] != '=')
1.70 rillig 448: goto maybe;
1.78 rillig 449: if (isalpha((unsigned char)*inp.s))
1.70 rillig 450: goto maybe;
451: return false;
452: maybe:
1.110 rillig 453: return ps.prev_token == lsym_semicolon ||
454: ps.prev_token == lsym_lbrace ||
455: ps.prev_token == lsym_rbrace;
1.57 rillig 456: }
457:
1.84 rillig 458: static int
459: bsearch_typenames(const char *key)
460: {
461: const char **arr = typenames.items;
462: int lo = 0;
463: int hi = (int)typenames.len - 1;
464:
465: while (lo <= hi) {
466: int mid = (int)((unsigned)(lo + hi) >> 1);
467: int cmp = strcmp(arr[mid], key);
468: if (cmp < 0)
469: lo = mid + 1;
470: else if (cmp > 0)
471: hi = mid - 1;
472: else
473: return mid;
474: }
475: return -(lo + 1);
476: }
477:
1.63 rillig 478: static bool
479: is_typename(void)
480: {
1.84 rillig 481: if (opt.auto_typedefs &&
482: token.e - token.s >= 2 && memcmp(token.e - 2, "_t", 2) == 0)
483: return true;
1.63 rillig 484:
1.84 rillig 485: return bsearch_typenames(token.s) >= 0;
1.63 rillig 486: }
487:
1.90 rillig 488: /* Read an alphanumeric token into 'token', or return end_of_file. */
1.100 rillig 489: static lexer_symbol
1.107 rillig 490: lexi_alnum(void)
1.1 cgd 491: {
1.89 rillig 492: if (isdigit((unsigned char)*inp.s) ||
493: (inp.s[0] == '.' && isdigit((unsigned char)inp.s[1]))) {
494: lex_number();
1.103 rillig 495: } else if (isalnum((unsigned char)*inp.s) ||
496: *inp.s == '_' || *inp.s == '$') {
1.89 rillig 497: lex_word();
1.102 rillig 498: } else
499: return lsym_eof; /* just as a placeholder */
500:
1.89 rillig 501: *token.e = '\0';
1.16 kamil 502:
1.89 rillig 503: if (token.s[0] == 'L' && token.s[1] == '\0' &&
504: (*inp.s == '"' || *inp.s == '\''))
1.100 rillig 505: return lsym_string_prefix;
1.16 kamil 506:
1.111 rillig 507: while (ch_isblank(inbuf_peek()))
1.32 rillig 508: inbuf_skip();
1.89 rillig 509:
1.110 rillig 510: if (ps.prev_token == lsym_tag && ps.p_l_follow == 0) {
1.107 rillig 511: ps.next_unary = true;
1.100 rillig 512: return lsym_type;
1.16 kamil 513: }
1.6 lukem 514:
1.89 rillig 515: /* Operator after identifier is binary unless last token was 'struct'. */
1.110 rillig 516: ps.next_unary = ps.prev_token == lsym_tag;
1.16 kamil 517:
1.89 rillig 518: const struct keyword *kw = bsearch(token.s, keywords,
1.94 rillig 519: array_length(keywords), sizeof(keywords[0]), cmp_keyword_by_name);
1.89 rillig 520: if (kw == NULL) {
521: if (is_typename()) {
1.107 rillig 522: ps.curr_keyword = kw_type;
523: ps.next_unary = true;
1.89 rillig 524: goto found_typename;
1.16 kamil 525: }
1.89 rillig 526:
527: } else { /* we have a keyword */
1.107 rillig 528: ps.curr_keyword = kw->kind;
529: ps.next_unary = true;
1.89 rillig 530:
531: switch (kw->kind) {
532: case kw_switch:
1.100 rillig 533: return lsym_switch;
1.89 rillig 534:
535: case kw_case_or_default:
1.100 rillig 536: return lsym_case_label;
1.89 rillig 537:
538: case kw_struct_or_union_or_enum:
539: case kw_type:
540: found_typename:
1.109 rillig 541: if (ps.p_l_follow > 0) {
1.108 rillig 542: /* inside parentheses: cast, param list, offsetof or sizeof */
1.107 rillig 543: ps.cast_mask |= (1 << ps.p_l_follow) & ~ps.not_cast_mask;
1.89 rillig 544: }
1.110 rillig 545: if (ps.prev_token == lsym_period ||
546: ps.prev_token == lsym_unary_op)
1.89 rillig 547: break;
548: if (kw != NULL && kw->kind == kw_struct_or_union_or_enum)
1.100 rillig 549: return lsym_tag;
1.109 rillig 550: if (ps.p_l_follow > 0)
1.89 rillig 551: break;
1.100 rillig 552: return lsym_type;
1.75 rillig 553:
1.98 rillig 554: case kw_for:
1.100 rillig 555: return lsym_for;
1.98 rillig 556:
557: case kw_if:
1.100 rillig 558: return lsym_if;
1.98 rillig 559:
560: case kw_while:
1.100 rillig 561: return lsym_while;
1.75 rillig 562:
1.97 rillig 563: case kw_do:
1.100 rillig 564: return lsym_do;
1.97 rillig 565:
566: case kw_else:
1.100 rillig 567: return lsym_else;
1.16 kamil 568:
1.89 rillig 569: case kw_storage_class:
1.100 rillig 570: return lsym_storage_class;
1.16 kamil 571:
1.89 rillig 572: case kw_typedef:
1.100 rillig 573: return lsym_typedef;
1.16 kamil 574:
1.89 rillig 575: default: /* all others are treated like any other
1.16 kamil 576: * identifier */
1.100 rillig 577: return lsym_ident;
1.90 rillig 578: }
579: }
1.89 rillig 580:
1.107 rillig 581: if (*inp.s == '(' && ps.tos <= 1 && ps.ind_level == 0 &&
582: !ps.in_parameter_declaration && !ps.block_init) {
1.89 rillig 583:
584: for (const char *p = inp.s; p < inp.e;)
585: if (*p++ == ')' && (*p == ';' || *p == ','))
586: goto not_proc;
587:
1.107 rillig 588: strncpy(ps.procname, token.s, sizeof ps.procname - 1);
589: if (ps.in_decl)
590: ps.in_parameter_declaration = true;
1.100 rillig 591: return lsym_funcname;
1.89 rillig 592: not_proc:;
593:
1.107 rillig 594: } else if (probably_typename()) {
595: ps.curr_keyword = kw_type;
596: ps.next_unary = true;
1.100 rillig 597: return lsym_type;
1.89 rillig 598: }
599:
1.110 rillig 600: if (ps.prev_token == lsym_type) /* if this is a declared variable,
1.89 rillig 601: * then following sign is unary */
1.107 rillig 602: ps.next_unary = true; /* will make "int a -1" work */
1.89 rillig 603:
1.100 rillig 604: return lsym_ident; /* the ident is not in the list */
1.89 rillig 605: }
1.75 rillig 606:
1.89 rillig 607: /* Reads the next token, placing it in the global variable "token". */
1.100 rillig 608: lexer_symbol
1.106 rillig 609: lexi(void)
1.89 rillig 610: {
1.90 rillig 611: token.e = token.s;
1.110 rillig 612: ps.prev_col_1 = ps.prev_newline;
613: ps.prev_newline = false;
1.107 rillig 614: ps.prev_keyword = ps.curr_keyword;
615: ps.curr_keyword = kw_0;
1.75 rillig 616:
1.111 rillig 617: while (ch_isblank(*inp.s)) {
1.110 rillig 618: ps.prev_col_1 = false;
1.89 rillig 619: inbuf_skip();
620: }
1.75 rillig 621:
1.107 rillig 622: lexer_symbol alnum_lsym = lexi_alnum();
1.100 rillig 623: if (alnum_lsym != lsym_eof)
1.107 rillig 624: return lexi_end(alnum_lsym);
1.16 kamil 625:
626: /* Scan a non-alphanumeric token */
627:
1.90 rillig 628: check_size_token(3); /* for things like "<<=" */
629: *token.e++ = inbuf_next();
1.50 rillig 630: *token.e = '\0';
1.16 kamil 631:
1.100 rillig 632: lexer_symbol lsym;
1.89 rillig 633: bool unary_delim = false; /* whether the current token forces a
634: * following operator to be unary */
635:
1.50 rillig 636: switch (*token.s) {
1.16 kamil 637: case '\n':
1.107 rillig 638: unary_delim = ps.next_unary;
1.110 rillig 639: ps.prev_newline = true;
1.47 rillig 640: /* if data has been exhausted, the newline is a dummy. */
1.100 rillig 641: lsym = had_eof ? lsym_eof : lsym_newline;
1.16 kamil 642: break;
643:
1.43 rillig 644: case '\'':
645: case '"':
1.44 rillig 646: lex_char_or_string();
1.100 rillig 647: lsym = lsym_ident;
1.16 kamil 648: break;
1.6 lukem 649:
1.40 rillig 650: case '(':
651: case '[':
1.16 kamil 652: unary_delim = true;
1.100 rillig 653: lsym = lsym_lparen_or_lbracket;
1.16 kamil 654: break;
655:
1.40 rillig 656: case ')':
657: case ']':
1.100 rillig 658: lsym = lsym_rparen_or_rbracket;
1.16 kamil 659: break;
660:
661: case '#':
1.107 rillig 662: unary_delim = ps.next_unary;
1.100 rillig 663: lsym = lsym_preprocessing;
1.16 kamil 664: break;
665:
666: case '?':
667: unary_delim = true;
1.100 rillig 668: lsym = lsym_question;
1.16 kamil 669: break;
670:
1.40 rillig 671: case ':':
1.100 rillig 672: lsym = lsym_colon;
1.16 kamil 673: unary_delim = true;
674: break;
675:
1.40 rillig 676: case ';':
1.16 kamil 677: unary_delim = true;
1.100 rillig 678: lsym = lsym_semicolon;
1.16 kamil 679: break;
680:
1.40 rillig 681: case '{':
1.16 kamil 682: unary_delim = true;
1.100 rillig 683: lsym = lsym_lbrace;
1.16 kamil 684: break;
685:
1.40 rillig 686: case '}':
1.16 kamil 687: unary_delim = true;
1.100 rillig 688: lsym = lsym_rbrace;
1.16 kamil 689: break;
690:
1.69 rillig 691: case '\f':
1.107 rillig 692: unary_delim = ps.next_unary;
1.110 rillig 693: ps.prev_newline = true;
1.100 rillig 694: lsym = lsym_form_feed;
1.16 kamil 695: break;
696:
1.40 rillig 697: case ',':
1.16 kamil 698: unary_delim = true;
1.100 rillig 699: lsym = lsym_comma;
1.16 kamil 700: break;
701:
702: case '.':
703: unary_delim = false;
1.100 rillig 704: lsym = lsym_period;
1.16 kamil 705: break;
1.1 cgd 706:
1.16 kamil 707: case '-':
1.90 rillig 708: case '+':
1.107 rillig 709: lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
1.16 kamil 710: unary_delim = true;
711:
1.90 rillig 712: if (*inp.s == token.s[0]) { /* ++, -- */
1.78 rillig 713: *token.e++ = *inp.s++;
1.110 rillig 714: if (ps.prev_token == lsym_ident ||
715: ps.prev_token == lsym_rparen_or_rbracket) {
1.107 rillig 716: lsym = ps.next_unary ? lsym_unary_op : lsym_postfix_op;
1.1 cgd 717: unary_delim = false;
1.16 kamil 718: }
1.75 rillig 719:
1.90 rillig 720: } else if (*inp.s == '=') { /* += */
1.78 rillig 721: *token.e++ = *inp.s++;
1.75 rillig 722:
1.90 rillig 723: } else if (*inp.s == '>') { /* -> */
1.78 rillig 724: *token.e++ = *inp.s++;
1.16 kamil 725: unary_delim = false;
1.100 rillig 726: lsym = lsym_unary_op;
1.107 rillig 727: ps.want_blank = false;
1.16 kamil 728: }
1.90 rillig 729: break;
1.16 kamil 730:
731: case '=':
1.107 rillig 732: if (ps.init_or_struct)
733: ps.block_init = true;
1.78 rillig 734: if (*inp.s == '=') { /* == */
735: *token.e++ = *inp.s++;
1.67 rillig 736: *token.e = '\0';
1.16 kamil 737: }
1.100 rillig 738: lsym = lsym_binary_op;
1.16 kamil 739: unary_delim = true;
740: break;
741:
742: case '>':
743: case '<':
744: case '!': /* ops like <, <<, <=, !=, etc */
1.78 rillig 745: if (*inp.s == '>' || *inp.s == '<' || *inp.s == '=')
1.50 rillig 746: *token.e++ = inbuf_next();
1.78 rillig 747: if (*inp.s == '=')
748: *token.e++ = *inp.s++;
1.107 rillig 749: lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
1.16 kamil 750: unary_delim = true;
751: break;
752:
753: case '*':
754: unary_delim = true;
1.107 rillig 755: if (!ps.next_unary) {
1.78 rillig 756: if (*inp.s == '=')
757: *token.e++ = *inp.s++;
1.100 rillig 758: lsym = lsym_binary_op;
1.16 kamil 759: break;
760: }
1.75 rillig 761:
1.78 rillig 762: while (*inp.s == '*' || isspace((unsigned char)*inp.s)) {
1.87 rillig 763: if (*inp.s == '*')
764: token_add_char('*');
1.32 rillig 765: inbuf_skip();
1.16 kamil 766: }
1.75 rillig 767:
1.16 kamil 768: if (ps.in_decl) {
1.78 rillig 769: char *tp = inp.s;
1.6 lukem 770:
1.16 kamil 771: while (isalpha((unsigned char)*tp) ||
1.103 rillig 772: isspace((unsigned char)*tp)) {
1.78 rillig 773: if (++tp >= inp.e)
1.81 rillig 774: inbuf_read_line();
1.16 kamil 775: }
776: if (*tp == '(')
777: ps.procname[0] = ' ';
778: }
1.75 rillig 779:
1.100 rillig 780: lsym = lsym_unary_op;
1.16 kamil 781: break;
1.1 cgd 782:
1.16 kamil 783: default:
1.78 rillig 784: if (token.s[0] == '/' && (*inp.s == '*' || *inp.s == '/')) {
1.16 kamil 785: /* it is start of comment */
1.50 rillig 786: *token.e++ = inbuf_next();
1.1 cgd 787:
1.100 rillig 788: lsym = lsym_comment;
1.107 rillig 789: unary_delim = ps.next_unary;
1.16 kamil 790: break;
1.1 cgd 791: }
1.75 rillig 792:
1.78 rillig 793: while (token.e[-1] == *inp.s || *inp.s == '=') {
1.87 rillig 794: /* handle '||', '&&', etc., and also things as in 'int *****i' */
795: token_add_char(inbuf_next());
1.16 kamil 796: }
1.75 rillig 797:
1.107 rillig 798: lsym = ps.next_unary ? lsym_unary_op : lsym_binary_op;
1.16 kamil 799: unary_delim = true;
1.47 rillig 800: }
1.16 kamil 801:
1.95 rillig 802: if (inp.s >= inp.e) /* check for input buffer empty */
1.81 rillig 803: inbuf_read_line();
1.75 rillig 804:
1.107 rillig 805: ps.next_unary = unary_delim;
1.75 rillig 806:
1.25 rillig 807: check_size_token(1);
1.50 rillig 808: *token.e = '\0';
1.75 rillig 809:
1.107 rillig 810: return lexi_end(lsym);
1.1 cgd 811: }
1.16 kamil 812:
1.6 lukem 813: void
1.64 rillig 814: add_typename(const char *name)
1.1 cgd 815: {
1.64 rillig 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: }
1.16 kamil 821:
1.84 rillig 822: int pos = bsearch_typenames(name);
1.64 rillig 823: if (pos >= 0)
824: return; /* already in the list */
1.75 rillig 825:
1.64 rillig 826: pos = -(pos + 1);
827: memmove(typenames.items + pos + 1, typenames.items + pos,
1.73 rillig 828: sizeof(typenames.items[0]) * (typenames.len++ - (unsigned)pos));
1.64 rillig 829: typenames.items[pos] = xstrdup(name);
1.1 cgd 830: }
CVSweb <webmaster@jp.NetBSD.org>