3 * Scanner for C declarations.
4 * Copyright © 2011, 2021, 2023 Nick Bowler
6 * This program is free software: you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation, either version 3 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License
17 * along with this program. If not, see <http://www.gnu.org/licenses/>.
27 %option nodefault noyywrap bison-locations reentrant never-interactive
28 %option extra-type="int"
29 %option prefix="cdecl__yy"
33 #include "cdecl-internal.h"
38 static char *to_octal(char *dst, unsigned val)
42 for (i = 0; i < 3; i++) {
43 *dst++ = '0' + ((val >> 6) & 7u);
51 * Convert a single character to a C-style character constant, including quote
52 * characters. At most 7 bytes are written to the buffer for the longest
53 * octal encoding, e.g., '\177'
55 static void to_readable_ch(char *dst, char c)
62 * The 7 standard C control characters are contiguous in ASCII,
63 * permitting a simple and compact lookup table; separating their
64 * handling from backslash and quote characters hopefully allows
65 * the compiler to recognize that.
68 case '\a': i = 0; break;
69 case '\b': i = 1; break;
70 case '\t': i = 2; break;
71 case '\n': i = 3; break;
72 case '\v': i = 4; break;
73 case '\f': i = 5; break;
74 case '\r': i = 6; break;
75 default: i = 7; break;
79 /* Otherwise printable characters that should still be escaped. */
81 case '\\': case '\'': esc = c; break;
88 } else if (isprint(uc)) {
92 dst = to_octal(dst, uc);
100 IDENT [_[:alpha:]][-_[:alnum:]]*
110 unsigned char *match;
111 static const unsigned char tab[2][8] = {
114 PACK_TOKEN(T_ASTERISK),
115 PACK_TOKEN(T_LBRACKET),
116 PACK_TOKEN(T_RBRACKET),
117 PACK_TOKEN(T_LPAREN),
118 PACK_TOKEN(T_RPAREN),
120 PACK_TOKEN(T_ELLIPSIS),
121 PACK_TOKEN(T_SEMICOLON)
125 match = memchr(&tab, yytext[0], sizeof tab[0]);
126 return UNPACK_TOKEN(match[sizeof tab[0]]);
129 0[0-7]* { intconv_base = INTCONV_OCTAL; goto int_parse; }
130 [1-9][0-9]* { intconv_base = INTCONV_DECIMAL; goto int_parse; }
136 intconv_base = INTCONV_HEXADECIMAL;
138 for (v = 0; (d = *yytext++);) {
139 if (!intconv_shift(&v, intconv_base, intconv_digit(d))) {
140 cdecl__errmsg(CDECL__ERANGE);
149 cdecl__errmsg(CDECL__EBADINT);
154 int len = yyleng, tok;
157 x = cdecl__to_keyword(yytext, len, yyextra);
158 yylval->spectype = UNPACK_SPEC(x & 0xff);
159 if ((tok = (x >> 8)) == PACK_TOKEN(T_IDENT)) {
161 * Our IDENT pattern includes hyphens so we can match
162 * "variable-length" as a keyword. In all other cases a
163 * hyphen is an error.
165 * We could use yyless to re-scan the hyphen and hit the
166 * error catch-all, but jumping straight to the error code
167 * seems to produce better results with gcc with no obvious
171 if ((c = memchr(yytext, '-', len)))
174 yyless(strcspn(yytext, "-"));
176 if (!(yylval->item = cdecl__alloc_item(len+1)))
178 memcpy(yylval->item->s, yytext, len+1);
180 return UNPACK_TOKEN(tok);
189 to_readable_ch(buf, *c);
190 cdecl__err(CDECL_ENOPARSE, _("syntax error, unexpected %s"), buf);