| @@ -1,1495 +1,1494 @@ | | | @@ -1,1495 +1,1494 @@ |
1 | /* $NetBSD: lex.c,v 1.155 2023/03/31 13:03:05 rillig Exp $ */ | | 1 | /* $NetBSD: lex.c,v 1.156 2023/04/05 20:17:30 rillig Exp $ */ |
2 | | | 2 | |
3 | /* | | 3 | /* |
4 | * Copyright (c) 1996 Christopher G. Demetriou. All Rights Reserved. | | 4 | * Copyright (c) 1996 Christopher G. Demetriou. All Rights Reserved. |
5 | * Copyright (c) 1994, 1995 Jochen Pohl | | 5 | * Copyright (c) 1994, 1995 Jochen Pohl |
6 | * All Rights Reserved. | | 6 | * All Rights Reserved. |
7 | * | | 7 | * |
8 | * Redistribution and use in source and binary forms, with or without | | 8 | * Redistribution and use in source and binary forms, with or without |
9 | * modification, are permitted provided that the following conditions | | 9 | * modification, are permitted provided that the following conditions |
10 | * are met: | | 10 | * are met: |
11 | * 1. Redistributions of source code must retain the above copyright | | 11 | * 1. Redistributions of source code must retain the above copyright |
12 | * notice, this list of conditions and the following disclaimer. | | 12 | * notice, this list of conditions and the following disclaimer. |
13 | * 2. Redistributions in binary form must reproduce the above copyright | | 13 | * 2. Redistributions in binary form must reproduce the above copyright |
14 | * notice, this list of conditions and the following disclaimer in the | | 14 | * notice, this list of conditions and the following disclaimer in the |
15 | * documentation and/or other materials provided with the distribution. | | 15 | * documentation and/or other materials provided with the distribution. |
16 | * 3. All advertising materials mentioning features or use of this software | | 16 | * 3. All advertising materials mentioning features or use of this software |
17 | * must display the following acknowledgement: | | 17 | * must display the following acknowledgement: |
18 | * This product includes software developed by Jochen Pohl for | | 18 | * This product includes software developed by Jochen Pohl for |
19 | * The NetBSD Project. | | 19 | * The NetBSD Project. |
20 | * 4. The name of the author may not be used to endorse or promote products | | 20 | * 4. The name of the author may not be used to endorse or promote products |
21 | * derived from this software without specific prior written permission. | | 21 | * derived from this software without specific prior written permission. |
22 | * | | 22 | * |
23 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR | | 23 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR |
24 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES | | 24 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES |
25 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. | | 25 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. |
26 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, | | 26 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, |
27 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT | | 27 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT |
28 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, | | 28 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
29 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY | | 29 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
30 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | | 30 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
31 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF | | 31 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF |
32 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | | 32 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
33 | */ | | 33 | */ |
34 | | | 34 | |
35 | #if HAVE_NBTOOL_CONFIG_H | | 35 | #if HAVE_NBTOOL_CONFIG_H |
36 | #include "nbtool_config.h" | | 36 | #include "nbtool_config.h" |
37 | #endif | | 37 | #endif |
38 | | | 38 | |
39 | #include <sys/cdefs.h> | | 39 | #include <sys/cdefs.h> |
40 | #if defined(__RCSID) | | 40 | #if defined(__RCSID) |
41 | __RCSID("$NetBSD: lex.c,v 1.155 2023/03/31 13:03:05 rillig Exp $"); | | 41 | __RCSID("$NetBSD: lex.c,v 1.156 2023/04/05 20:17:30 rillig Exp $"); |
42 | #endif | | 42 | #endif |
43 | | | 43 | |
44 | #include <ctype.h> | | 44 | #include <ctype.h> |
45 | #include <errno.h> | | 45 | #include <errno.h> |
46 | #include <float.h> | | 46 | #include <float.h> |
47 | #include <limits.h> | | 47 | #include <limits.h> |
48 | #include <math.h> | | 48 | #include <math.h> |
49 | #include <stdlib.h> | | 49 | #include <stdlib.h> |
50 | #include <string.h> | | 50 | #include <string.h> |
51 | | | 51 | |
52 | #include "lint1.h" | | 52 | #include "lint1.h" |
53 | #include "cgram.h" | | 53 | #include "cgram.h" |
54 | | | 54 | |
55 | #define CHAR_MASK ((1U << CHAR_SIZE) - 1) | | 55 | #define CHAR_MASK ((1U << CHAR_SIZE) - 1) |
56 | | | 56 | |
57 | | | 57 | |
58 | /* Current position (it's also updated when an included file is parsed) */ | | 58 | /* Current position (it's also updated when an included file is parsed) */ |
59 | pos_t curr_pos = { "", 1, 0 }; | | 59 | pos_t curr_pos = { "", 1, 0 }; |
60 | | | 60 | |
61 | /* | | 61 | /* |
62 | * Current position in C source (not updated when an included file is | | 62 | * Current position in C source (not updated when an included file is |
63 | * parsed). | | 63 | * parsed). |
64 | */ | | 64 | */ |
65 | pos_t csrc_pos = { "", 1, 0 }; | | 65 | pos_t csrc_pos = { "", 1, 0 }; |
66 | | | 66 | |
67 | bool in_gcc_attribute; | | 67 | bool in_gcc_attribute; |
68 | bool in_system_header; | | 68 | bool in_system_header; |
69 | | | 69 | |
70 | /* | | 70 | /* |
71 | * Valid values for 'since' are 78, 90, 99, 11. | | 71 | * Valid values for 'since' are 78, 90, 99, 11. |
72 | * | | 72 | * |
73 | * The C11 keywords are added in C99 mode as well, to provide good error | | 73 | * The C11 keywords are added in C99 mode as well, to provide good error |
74 | * messages instead of a simple parse error. If the keyword '_Generic' were | | 74 | * messages instead of a simple parse error. If the keyword '_Generic' were |
75 | * not defined, it would be interpreted as an implicit function call, leading | | 75 | * not defined, it would be interpreted as an implicit function call, leading |
76 | * to a parse error. | | 76 | * to a parse error. |
77 | */ | | 77 | */ |
78 | #define kwdef(name, token, scl, tspec, tqual, since, gcc, deco) \ | | 78 | #define kwdef(name, token, scl, tspec, tqual, since, gcc, deco) \ |
79 | { \ | | 79 | { \ |
80 | name, token, scl, tspec, tqual, \ | | 80 | name, token, scl, tspec, tqual, \ |
81 | (since) == 90, \ | | 81 | (since) == 90, \ |
82 | /* CONSTCOND */ (since) == 99 || (since) == 11, \ | | 82 | /* CONSTCOND */ (since) == 99 || (since) == 11, \ |
83 | (gcc) > 0, \ | | 83 | (gcc) > 0, \ |
84 | ((deco) & 1) != 0, ((deco) & 2) != 0, ((deco) & 4) != 0, \ | | 84 | ((deco) & 1) != 0, ((deco) & 2) != 0, ((deco) & 4) != 0, \ |
85 | } | | 85 | } |
86 | #define kwdef_token(name, token, since, gcc, deco) \ | | 86 | #define kwdef_token(name, token, since, gcc, deco) \ |
87 | kwdef(name, token, 0, 0, 0, since, gcc, deco) | | 87 | kwdef(name, token, 0, 0, 0, since, gcc, deco) |
88 | #define kwdef_sclass(name, sclass, since, gcc, deco) \ | | 88 | #define kwdef_sclass(name, sclass, since, gcc, deco) \ |
89 | kwdef(name, T_SCLASS, sclass, 0, 0, since, gcc, deco) | | 89 | kwdef(name, T_SCLASS, sclass, 0, 0, since, gcc, deco) |
90 | #define kwdef_type(name, tspec, since) \ | | 90 | #define kwdef_type(name, tspec, since) \ |
91 | kwdef(name, T_TYPE, 0, tspec, 0, since, 0, 1) | | 91 | kwdef(name, T_TYPE, 0, tspec, 0, since, 0, 1) |
92 | #define kwdef_tqual(name, tqual, since, gcc, deco) \ | | 92 | #define kwdef_tqual(name, tqual, since, gcc, deco) \ |
93 | kwdef(name, T_QUAL, 0, 0, tqual, since, gcc, deco) | | 93 | kwdef(name, T_QUAL, 0, 0, tqual, since, gcc, deco) |
94 | #define kwdef_keyword(name, token) \ | | 94 | #define kwdef_keyword(name, token) \ |
95 | kwdef(name, token, 0, 0, 0, 78, 0, 1) | | 95 | kwdef(name, token, 0, 0, 0, 78, 0, 1) |
96 | | | 96 | |
97 | /* During initialization, these keywords are written to the symbol table. */ | | 97 | /* During initialization, these keywords are written to the symbol table. */ |
98 | static const struct keyword { | | 98 | static const struct keyword { |
99 | const char *kw_name; | | 99 | const char *kw_name; |
100 | int kw_token; /* token returned by yylex() */ | | 100 | int kw_token; /* token returned by yylex() */ |
101 | scl_t kw_scl; /* storage class if kw_token is T_SCLASS */ | | 101 | scl_t kw_scl; /* storage class if kw_token is T_SCLASS */ |
102 | tspec_t kw_tspec; /* type specifier if kw_token is T_TYPE or | | 102 | tspec_t kw_tspec; /* type specifier if kw_token is T_TYPE or |
103 | * T_STRUCT_OR_UNION */ | | 103 | * T_STRUCT_OR_UNION */ |
104 | tqual_t kw_tqual; /* type qualifier if kw_token is T_QUAL */ | | 104 | tqual_t kw_tqual; /* type qualifier if kw_token is T_QUAL */ |
105 | bool kw_c90:1; /* available in C90 mode */ | | 105 | bool kw_c90:1; /* available in C90 mode */ |
106 | bool kw_c99_or_c11:1; /* available in C99 or C11 mode */ | | 106 | bool kw_c99_or_c11:1; /* available in C99 or C11 mode */ |
107 | bool kw_gcc:1; /* available in GCC mode */ | | 107 | bool kw_gcc:1; /* available in GCC mode */ |
108 | bool kw_plain:1; /* 'name' */ | | 108 | bool kw_plain:1; /* 'name' */ |
109 | bool kw_leading:1; /* '__name' */ | | 109 | bool kw_leading:1; /* '__name' */ |
110 | bool kw_both:1; /* '__name__' */ | | 110 | bool kw_both:1; /* '__name__' */ |
111 | } keywords[] = { | | 111 | } keywords[] = { |
112 | kwdef_keyword( "_Alignas", T_ALIGNAS), | | 112 | kwdef_keyword( "_Alignas", T_ALIGNAS), |
113 | kwdef_keyword( "_Alignof", T_ALIGNOF), | | 113 | kwdef_keyword( "_Alignof", T_ALIGNOF), |
114 | kwdef_token( "alignof", T_ALIGNOF, 78,0,6), | | 114 | kwdef_token( "alignof", T_ALIGNOF, 78,0,6), |
115 | kwdef_token( "asm", T_ASM, 78,1,7), | | 115 | kwdef_token( "asm", T_ASM, 78,1,7), |
116 | kwdef_token( "_Atomic", T_ATOMIC, 11,0,1), | | 116 | kwdef_token( "_Atomic", T_ATOMIC, 11,0,1), |
117 | kwdef_token( "attribute", T_ATTRIBUTE, 78,1,6), | | 117 | kwdef_token( "attribute", T_ATTRIBUTE, 78,1,6), |
118 | kwdef_sclass( "auto", AUTO, 78,0,1), | | 118 | kwdef_sclass( "auto", AUTO, 78,0,1), |
119 | kwdef_type( "_Bool", BOOL, 99), | | 119 | kwdef_type( "_Bool", BOOL, 99), |
120 | kwdef_keyword( "break", T_BREAK), | | 120 | kwdef_keyword( "break", T_BREAK), |
121 | kwdef_token( "__builtin_offsetof", T_BUILTIN_OFFSETOF, 78,1,1), | | 121 | kwdef_token( "__builtin_offsetof", T_BUILTIN_OFFSETOF, 78,1,1), |
122 | kwdef_keyword( "case", T_CASE), | | 122 | kwdef_keyword( "case", T_CASE), |
123 | kwdef_type( "char", CHAR, 78), | | 123 | kwdef_type( "char", CHAR, 78), |
124 | kwdef_type( "_Complex", COMPLEX, 99), | | 124 | kwdef_type( "_Complex", COMPLEX, 99), |
125 | kwdef_tqual( "const", CONST, 90,0,7), | | 125 | kwdef_tqual( "const", CONST, 90,0,7), |
126 | kwdef_keyword( "continue", T_CONTINUE), | | 126 | kwdef_keyword( "continue", T_CONTINUE), |
127 | kwdef_keyword( "default", T_DEFAULT), | | 127 | kwdef_keyword( "default", T_DEFAULT), |
128 | kwdef_keyword( "do", T_DO), | | 128 | kwdef_keyword( "do", T_DO), |
129 | kwdef_type( "double", DOUBLE, 78), | | 129 | kwdef_type( "double", DOUBLE, 78), |
130 | kwdef_keyword( "else", T_ELSE), | | 130 | kwdef_keyword( "else", T_ELSE), |
131 | kwdef_keyword( "enum", T_ENUM), | | 131 | kwdef_keyword( "enum", T_ENUM), |
132 | kwdef_token( "__extension__",T_EXTENSION, 78,1,1), | | 132 | kwdef_token( "__extension__",T_EXTENSION, 78,1,1), |
133 | kwdef_sclass( "extern", EXTERN, 78,0,1), | | 133 | kwdef_sclass( "extern", EXTERN, 78,0,1), |
134 | kwdef_type( "float", FLOAT, 78), | | 134 | kwdef_type( "float", FLOAT, 78), |
135 | kwdef_keyword( "for", T_FOR), | | 135 | kwdef_keyword( "for", T_FOR), |
136 | kwdef_token( "_Generic", T_GENERIC, 11,0,1), | | 136 | kwdef_token( "_Generic", T_GENERIC, 11,0,1), |
137 | kwdef_keyword( "goto", T_GOTO), | | 137 | kwdef_keyword( "goto", T_GOTO), |
138 | kwdef_keyword( "if", T_IF), | | 138 | kwdef_keyword( "if", T_IF), |
139 | kwdef_token( "__imag__", T_IMAG, 78,1,1), | | 139 | kwdef_token( "__imag__", T_IMAG, 78,1,1), |
140 | kwdef_sclass( "inline", INLINE, 99,0,7), | | 140 | kwdef_sclass( "inline", INLINE, 99,0,7), |
141 | kwdef_type( "int", INT, 78), | | 141 | kwdef_type( "int", INT, 78), |
142 | #ifdef INT128_SIZE | | 142 | #ifdef INT128_SIZE |
143 | kwdef_type( "__int128_t", INT128, 99), | | 143 | kwdef_type( "__int128_t", INT128, 99), |
144 | #endif | | 144 | #endif |
145 | kwdef_type( "long", LONG, 78), | | 145 | kwdef_type( "long", LONG, 78), |
146 | kwdef_token( "_Noreturn", T_NORETURN, 11,0,1), | | 146 | kwdef_token( "_Noreturn", T_NORETURN, 11,0,1), |
147 | kwdef_token( "__packed", T_PACKED, 78,0,1), | | 147 | kwdef_token( "__packed", T_PACKED, 78,0,1), |
148 | kwdef_token( "__real__", T_REAL, 78,1,1), | | 148 | kwdef_token( "__real__", T_REAL, 78,1,1), |
149 | kwdef_sclass( "register", REG, 78,0,1), | | 149 | kwdef_sclass( "register", REG, 78,0,1), |
150 | kwdef_tqual( "restrict", RESTRICT, 99,0,7), | | 150 | kwdef_tqual( "restrict", RESTRICT, 99,0,7), |
151 | kwdef_keyword( "return", T_RETURN), | | 151 | kwdef_keyword( "return", T_RETURN), |
152 | kwdef_type( "short", SHORT, 78), | | 152 | kwdef_type( "short", SHORT, 78), |
153 | kwdef( "signed", T_TYPE, 0, SIGNED, 0, 90,0,3), | | 153 | kwdef( "signed", T_TYPE, 0, SIGNED, 0, 90,0,3), |
154 | kwdef_keyword( "sizeof", T_SIZEOF), | | 154 | kwdef_keyword( "sizeof", T_SIZEOF), |
155 | kwdef_sclass( "static", STATIC, 78,0,1), | | 155 | kwdef_sclass( "static", STATIC, 78,0,1), |
156 | kwdef_keyword( "_Static_assert", T_STATIC_ASSERT), | | 156 | kwdef_keyword( "_Static_assert", T_STATIC_ASSERT), |
157 | kwdef("struct", T_STRUCT_OR_UNION, 0, STRUCT, 0, 78,0,1), | | 157 | kwdef("struct", T_STRUCT_OR_UNION, 0, STRUCT, 0, 78,0,1), |
158 | kwdef_keyword( "switch", T_SWITCH), | | 158 | kwdef_keyword( "switch", T_SWITCH), |
159 | kwdef_token( "__symbolrename", T_SYMBOLRENAME, 78,0,1), | | 159 | kwdef_token( "__symbolrename", T_SYMBOLRENAME, 78,0,1), |
160 | kwdef_tqual( "__thread", THREAD, 78,1,1), | | 160 | kwdef_tqual( "__thread", THREAD, 78,1,1), |
161 | /* XXX: _Thread_local is a storage-class-specifier, not tqual. */ | | 161 | /* XXX: _Thread_local is a storage-class-specifier, not tqual. */ |
162 | kwdef_tqual( "_Thread_local", THREAD, 11,0,1), | | 162 | kwdef_tqual( "_Thread_local", THREAD, 11,0,1), |
163 | kwdef_sclass( "typedef", TYPEDEF, 78,0,1), | | 163 | kwdef_sclass( "typedef", TYPEDEF, 78,0,1), |
164 | kwdef_token( "typeof", T_TYPEOF, 78,1,7), | | 164 | kwdef_token( "typeof", T_TYPEOF, 78,1,7), |
165 | #ifdef INT128_SIZE | | 165 | #ifdef INT128_SIZE |
166 | kwdef_type( "__uint128_t", UINT128, 99), | | 166 | kwdef_type( "__uint128_t", UINT128, 99), |
167 | #endif | | 167 | #endif |
168 | kwdef("union", T_STRUCT_OR_UNION, 0, UNION, 0, 78,0,1), | | 168 | kwdef("union", T_STRUCT_OR_UNION, 0, UNION, 0, 78,0,1), |
169 | kwdef_type( "unsigned", UNSIGN, 78), | | 169 | kwdef_type( "unsigned", UNSIGN, 78), |
170 | kwdef_type( "void", VOID, 78), | | 170 | kwdef_type( "void", VOID, 78), |
171 | kwdef_tqual( "volatile", VOLATILE, 90,0,7), | | 171 | kwdef_tqual( "volatile", VOLATILE, 90,0,7), |
172 | kwdef_keyword( "while", T_WHILE), | | 172 | kwdef_keyword( "while", T_WHILE), |
173 | #undef kwdef | | 173 | #undef kwdef |
174 | #undef kwdef_token | | 174 | #undef kwdef_token |
175 | #undef kwdef_sclass | | 175 | #undef kwdef_sclass |
176 | #undef kwdef_type | | 176 | #undef kwdef_type |
177 | #undef kwdef_tqual | | 177 | #undef kwdef_tqual |
178 | #undef kwdef_keyword | | 178 | #undef kwdef_keyword |
179 | }; | | 179 | }; |
180 | | | 180 | |
181 | /* | | 181 | /* |
182 | * The symbol table containing all keywords, identifiers and labels. The hash | | 182 | * The symbol table containing all keywords, identifiers and labels. The hash |
183 | * entries are linked via sym_t.s_symtab_next. | | 183 | * entries are linked via sym_t.s_symtab_next. |
184 | */ | | 184 | */ |
185 | static sym_t *symtab[HSHSIZ1]; | | 185 | static sym_t *symtab[HSHSIZ1]; |
186 | | | 186 | |
187 | /* | | 187 | /* |
188 | * The kind of the next expected symbol, to distinguish the namespaces of | | 188 | * The kind of the next expected symbol, to distinguish the namespaces of |
189 | * members, labels, type tags and other identifiers. | | 189 | * members, labels, type tags and other identifiers. |
190 | */ | | 190 | */ |
191 | symt_t symtyp; | | 191 | symt_t symtyp; |
192 | | | 192 | |
193 | | | 193 | |
194 | static unsigned int | | 194 | static unsigned int |
195 | hash(const char *s) | | 195 | hash(const char *s) |
196 | { | | 196 | { |
197 | unsigned int v; | | 197 | unsigned int v; |
198 | const char *p; | | 198 | const char *p; |
199 | | | 199 | |
200 | v = 0; | | 200 | v = 0; |
201 | for (p = s; *p != '\0'; p++) { | | 201 | for (p = s; *p != '\0'; p++) { |
202 | v = (v << 4) + (unsigned char)*p; | | 202 | v = (v << 4) + (unsigned char)*p; |
203 | v ^= v >> 28; | | 203 | v ^= v >> 28; |
204 | } | | 204 | } |
205 | return v % HSHSIZ1; | | 205 | return v % HSHSIZ1; |
206 | } | | 206 | } |
207 | | | 207 | |
208 | static void | | 208 | static void |
209 | symtab_add(sym_t *sym) | | 209 | symtab_add(sym_t *sym) |
210 | { | | 210 | { |
211 | unsigned int h; | | 211 | unsigned int h; |
212 | | | 212 | |
213 | h = hash(sym->s_name); | | 213 | h = hash(sym->s_name); |
214 | if ((sym->s_symtab_next = symtab[h]) != NULL) | | 214 | if ((sym->s_symtab_next = symtab[h]) != NULL) |
215 | symtab[h]->s_symtab_ref = &sym->s_symtab_next; | | 215 | symtab[h]->s_symtab_ref = &sym->s_symtab_next; |
216 | sym->s_symtab_ref = &symtab[h]; | | 216 | sym->s_symtab_ref = &symtab[h]; |
217 | symtab[h] = sym; | | 217 | symtab[h] = sym; |
218 | } | | 218 | } |
219 | | | 219 | |
220 | static sym_t * | | 220 | static sym_t * |
221 | symtab_search(const char *name) | | 221 | symtab_search(const char *name) |
222 | { | | 222 | { |
223 | | | 223 | |
224 | unsigned int h = hash(name); | | 224 | unsigned int h = hash(name); |
225 | for (sym_t *sym = symtab[h]; sym != NULL; sym = sym->s_symtab_next) { | | 225 | for (sym_t *sym = symtab[h]; sym != NULL; sym = sym->s_symtab_next) { |
226 | if (strcmp(sym->s_name, name) != 0) | | 226 | if (strcmp(sym->s_name, name) != 0) |
227 | continue; | | 227 | continue; |
228 | if (sym->s_keyword != NULL || | | 228 | if (sym->s_keyword != NULL || |
229 | sym->s_kind == symtyp || | | 229 | sym->s_kind == symtyp || |
230 | in_gcc_attribute) | | 230 | in_gcc_attribute) |
231 | return sym; | | 231 | return sym; |
232 | } | | 232 | } |
233 | | | 233 | |
234 | return NULL; | | 234 | return NULL; |
235 | } | | 235 | } |
236 | | | 236 | |
237 | static void | | 237 | static void |
238 | symtab_remove(sym_t *sym) | | 238 | symtab_remove(sym_t *sym) |
239 | { | | 239 | { |
240 | | | 240 | |
241 | if ((*sym->s_symtab_ref = sym->s_symtab_next) != NULL) | | 241 | if ((*sym->s_symtab_ref = sym->s_symtab_next) != NULL) |
242 | sym->s_symtab_next->s_symtab_ref = sym->s_symtab_ref; | | 242 | sym->s_symtab_next->s_symtab_ref = sym->s_symtab_ref; |
243 | sym->s_symtab_next = NULL; | | 243 | sym->s_symtab_next = NULL; |
244 | } | | 244 | } |
245 | | | 245 | |
246 | static void | | 246 | static void |
247 | symtab_remove_locals(void) | | 247 | symtab_remove_locals(void) |
248 | { | | 248 | { |
249 | | | 249 | |
250 | for (size_t i = 0; i < HSHSIZ1; i++) { | | 250 | for (size_t i = 0; i < HSHSIZ1; i++) { |
251 | for (sym_t *sym = symtab[i]; sym != NULL; ) { | | 251 | for (sym_t *sym = symtab[i]; sym != NULL; ) { |
252 | sym_t *next = sym->s_symtab_next; | | 252 | sym_t *next = sym->s_symtab_next; |
253 | if (sym->s_block_level >= 1) | | 253 | if (sym->s_block_level >= 1) |
254 | symtab_remove(sym); | | 254 | symtab_remove(sym); |
255 | sym = next; | | 255 | sym = next; |
256 | } | | 256 | } |
257 | } | | 257 | } |
258 | } | | 258 | } |
259 | | | 259 | |
260 | #ifdef DEBUG | | 260 | #ifdef DEBUG |
261 | static int | | 261 | static int |
262 | sym_by_name(const void *va, const void *vb) | | 262 | sym_by_name(const void *va, const void *vb) |
263 | { | | 263 | { |
264 | const sym_t *a = *(const sym_t *const *)va; | | 264 | const sym_t *a = *(const sym_t *const *)va; |
265 | const sym_t *b = *(const sym_t *const *)vb; | | 265 | const sym_t *b = *(const sym_t *const *)vb; |
266 | | | 266 | |
267 | return strcmp(a->s_name, b->s_name); | | 267 | return strcmp(a->s_name, b->s_name); |
268 | } | | 268 | } |
269 | | | 269 | |
270 | struct syms { | | 270 | struct syms { |
271 | const sym_t **items; | | 271 | const sym_t **items; |
272 | size_t len; | | 272 | size_t len; |
273 | size_t cap; | | 273 | size_t cap; |
274 | }; | | 274 | }; |
275 | | | 275 | |
276 | static void | | 276 | static void |
277 | syms_add(struct syms *syms, const sym_t *sym) | | 277 | syms_add(struct syms *syms, const sym_t *sym) |
278 | { | | 278 | { |
279 | if (syms->len >= syms->cap) { | | 279 | if (syms->len >= syms->cap) { |
280 | syms->cap *= 2; | | 280 | syms->cap *= 2; |
281 | syms->items = xrealloc(syms->items, | | 281 | syms->items = xrealloc(syms->items, |
282 | syms->cap * sizeof(syms->items[0])); | | 282 | syms->cap * sizeof(syms->items[0])); |
283 | } | | 283 | } |
284 | syms->items[syms->len++] = sym; | | 284 | syms->items[syms->len++] = sym; |
285 | } | | 285 | } |
286 | | | 286 | |
287 | void | | 287 | void |
288 | debug_symtab(void) | | 288 | debug_symtab(void) |
289 | { | | 289 | { |
290 | struct syms syms = { xcalloc(64, sizeof(syms.items[0])), 0, 64 }; | | 290 | struct syms syms = { xcalloc(64, sizeof(syms.items[0])), 0, 64 }; |
291 | | | 291 | |
292 | for (int level = -1;; level++) { | | 292 | for (int level = -1;; level++) { |
293 | bool more = false; | | 293 | bool more = false; |
294 | size_t n = sizeof(symtab) / sizeof(symtab[0]); | | 294 | size_t n = sizeof(symtab) / sizeof(symtab[0]); |
295 | | | 295 | |
296 | syms.len = 0; | | 296 | syms.len = 0; |
297 | for (size_t i = 0; i < n; i++) { | | 297 | for (size_t i = 0; i < n; i++) { |
298 | for (sym_t *sym = symtab[i]; sym != NULL;) { | | 298 | for (sym_t *sym = symtab[i]; sym != NULL;) { |
299 | if (sym->s_block_level == level && | | 299 | if (sym->s_block_level == level && |
300 | sym->s_keyword == NULL) | | 300 | sym->s_keyword == NULL) |
301 | syms_add(&syms, sym); | | 301 | syms_add(&syms, sym); |
302 | if (sym->s_block_level > level) | | 302 | if (sym->s_block_level > level) |
303 | more = true; | | 303 | more = true; |
304 | sym = sym->s_symtab_next; | | 304 | sym = sym->s_symtab_next; |
305 | } | | 305 | } |
306 | } | | 306 | } |
307 | | | 307 | |
308 | if (syms.len > 0) { | | 308 | if (syms.len > 0) { |
309 | debug_printf("symbol table level %d\n", level); | | 309 | debug_printf("symbol table level %d\n", level); |
310 | debug_indent_inc(); | | 310 | debug_indent_inc(); |
311 | qsort(syms.items, syms.len, sizeof(syms.items[0]), | | 311 | qsort(syms.items, syms.len, sizeof(syms.items[0]), |
312 | sym_by_name); | | 312 | sym_by_name); |
313 | for (size_t i = 0; i < syms.len; i++) | | 313 | for (size_t i = 0; i < syms.len; i++) |
314 | debug_sym("", syms.items[i], "\n"); | | 314 | debug_sym("", syms.items[i], "\n"); |
315 | debug_indent_dec(); | | 315 | debug_indent_dec(); |
316 | | | 316 | |
317 | lint_assert(level != -1); | | 317 | lint_assert(level != -1); |
318 | } | | 318 | } |
319 | | | 319 | |
320 | if (!more) | | 320 | if (!more) |
321 | break; | | 321 | break; |
322 | } | | 322 | } |
323 | | | 323 | |
324 | free(syms.items); | | 324 | free(syms.items); |
325 | } | | 325 | } |
326 | #endif | | 326 | #endif |
327 | | | 327 | |
328 | static void | | 328 | static void |
329 | add_keyword(const struct keyword *kw, bool leading, bool trailing) | | 329 | add_keyword(const struct keyword *kw, bool leading, bool trailing) |
330 | { | | 330 | { |
331 | | | 331 | |
332 | const char *name; | | 332 | const char *name; |
333 | if (!leading && !trailing) { | | 333 | if (!leading && !trailing) { |
334 | name = kw->kw_name; | | 334 | name = kw->kw_name; |
335 | } else { | | 335 | } else { |
336 | char buf[256]; | | 336 | char buf[256]; |
337 | (void)snprintf(buf, sizeof(buf), "%s%s%s", | | 337 | (void)snprintf(buf, sizeof(buf), "%s%s%s", |
338 | leading ? "__" : "", kw->kw_name, trailing ? "__" : ""); | | 338 | leading ? "__" : "", kw->kw_name, trailing ? "__" : ""); |
339 | name = xstrdup(buf); | | 339 | name = xstrdup(buf); |
340 | } | | 340 | } |
341 | | | 341 | |
342 | sym_t *sym = block_zero_alloc(sizeof(*sym)); | | 342 | sym_t *sym = block_zero_alloc(sizeof(*sym)); |
343 | sym->s_name = name; | | 343 | sym->s_name = name; |
344 | sym->s_keyword = kw; | | 344 | sym->s_keyword = kw; |
345 | int tok = kw->kw_token; | | 345 | int tok = kw->kw_token; |
346 | sym->u.s_keyword.sk_token = tok; | | 346 | sym->u.s_keyword.sk_token = tok; |
347 | if (tok == T_TYPE || tok == T_STRUCT_OR_UNION) | | 347 | if (tok == T_TYPE || tok == T_STRUCT_OR_UNION) |
348 | sym->u.s_keyword.sk_tspec = kw->kw_tspec; | | 348 | sym->u.s_keyword.sk_tspec = kw->kw_tspec; |
349 | if (tok == T_SCLASS) | | 349 | if (tok == T_SCLASS) |
350 | sym->s_scl = kw->kw_scl; | | 350 | sym->s_scl = kw->kw_scl; |
351 | if (tok == T_QUAL) | | 351 | if (tok == T_QUAL) |
352 | sym->u.s_keyword.sk_qualifier = kw->kw_tqual; | | 352 | sym->u.s_keyword.sk_qualifier = kw->kw_tqual; |
353 | | | 353 | |
354 | symtab_add(sym); | | 354 | symtab_add(sym); |
355 | } | | 355 | } |
356 | | | 356 | |
357 | static bool | | 357 | static bool |
358 | is_keyword_known(const struct keyword *kw) | | 358 | is_keyword_known(const struct keyword *kw) |
359 | { | | 359 | { |
360 | | | 360 | |
361 | if ((kw->kw_c90 || kw->kw_c99_or_c11) && !allow_c90) | | 361 | if ((kw->kw_c90 || kw->kw_c99_or_c11) && !allow_c90) |
362 | return false; | | 362 | return false; |
363 | | | 363 | |
364 | /* | | 364 | /* |
365 | * In the 1990s, GCC defined several keywords that were later | | 365 | * In the 1990s, GCC defined several keywords that were later |
366 | * incorporated into C99, therefore in GCC mode, all C99 keywords are | | 366 | * incorporated into C99, therefore in GCC mode, all C99 keywords are |
367 | * made available. The C11 keywords are made available as well, but | | 367 | * made available. The C11 keywords are made available as well, but |
368 | * there are so few that they don't matter practically. | | 368 | * there are so few that they don't matter practically. |
369 | */ | | 369 | */ |
370 | if (allow_gcc) | | 370 | if (allow_gcc) |
371 | return true; | | 371 | return true; |
372 | if (kw->kw_gcc) | | 372 | if (kw->kw_gcc) |
373 | return false; | | 373 | return false; |
374 | | | 374 | |
375 | if (kw->kw_c99_or_c11 && !allow_c99) | | 375 | if (kw->kw_c99_or_c11 && !allow_c99) |
376 | return false; | | 376 | return false; |
377 | return true; | | 377 | return true; |
378 | } | | 378 | } |
379 | | | 379 | |
380 | /* Write all keywords to the symbol table. */ | | 380 | /* Write all keywords to the symbol table. */ |
381 | void | | 381 | void |
382 | initscan(void) | | 382 | initscan(void) |
383 | { | | 383 | { |
384 | | | 384 | |
385 | size_t n = sizeof(keywords) / sizeof(keywords[0]); | | 385 | size_t n = sizeof(keywords) / sizeof(keywords[0]); |
386 | for (size_t i = 0; i < n; i++) { | | 386 | for (size_t i = 0; i < n; i++) { |
387 | const struct keyword *kw = keywords + i; | | 387 | const struct keyword *kw = keywords + i; |
388 | if (!is_keyword_known(kw)) | | 388 | if (!is_keyword_known(kw)) |
389 | continue; | | 389 | continue; |
390 | if (kw->kw_plain) | | 390 | if (kw->kw_plain) |
391 | add_keyword(kw, false, false); | | 391 | add_keyword(kw, false, false); |
392 | if (kw->kw_leading) | | 392 | if (kw->kw_leading) |
393 | add_keyword(kw, true, false); | | 393 | add_keyword(kw, true, false); |
394 | if (kw->kw_both) | | 394 | if (kw->kw_both) |
395 | add_keyword(kw, true, true); | | 395 | add_keyword(kw, true, true); |
396 | } | | 396 | } |
397 | } | | 397 | } |
398 | | | 398 | |
399 | /* | | 399 | /* |
400 | * When scanning the remainder of a long token (see lex_input), read a byte | | 400 | * When scanning the remainder of a long token (see lex_input), read a byte |
401 | * and return it as an unsigned char or as EOF. | | 401 | * and return it as an unsigned char or as EOF. |
402 | * | | 402 | * |
403 | * Increment the line counts if necessary. | | 403 | * Increment the line counts if necessary. |
404 | */ | | 404 | */ |
405 | static int | | 405 | static int |
406 | read_byte(void) | | 406 | read_byte(void) |
407 | { | | 407 | { |
408 | int c; | | 408 | int c; |
409 | | | 409 | |
410 | if ((c = lex_input()) == EOF) | | 410 | if ((c = lex_input()) == EOF) |
411 | return c; | | 411 | return c; |
412 | if (c == '\0') | | 412 | if (c == '\0') |
413 | return EOF; /* lex returns 0 on EOF. */ | | 413 | return EOF; /* lex returns 0 on EOF. */ |
414 | if (c == '\n') | | 414 | if (c == '\n') |
415 | lex_next_line(); | | 415 | lex_next_line(); |
416 | return c; | | 416 | return c; |
417 | } | | 417 | } |
418 | | | 418 | |
419 | static int | | 419 | static int |
420 | lex_keyword(sym_t *sym) | | 420 | lex_keyword(sym_t *sym) |
421 | { | | 421 | { |
422 | int tok = sym->u.s_keyword.sk_token; | | 422 | int tok = sym->u.s_keyword.sk_token; |
423 | | | 423 | |
424 | if (tok == T_SCLASS) | | 424 | if (tok == T_SCLASS) |
425 | yylval.y_scl = sym->s_scl; | | 425 | yylval.y_scl = sym->s_scl; |
426 | if (tok == T_TYPE || tok == T_STRUCT_OR_UNION) | | 426 | if (tok == T_TYPE || tok == T_STRUCT_OR_UNION) |
427 | yylval.y_tspec = sym->u.s_keyword.sk_tspec; | | 427 | yylval.y_tspec = sym->u.s_keyword.sk_tspec; |
428 | if (tok == T_QUAL) | | 428 | if (tok == T_QUAL) |
429 | yylval.y_tqual = sym->u.s_keyword.sk_qualifier; | | 429 | yylval.y_tqual = sym->u.s_keyword.sk_qualifier; |
430 | return tok; | | 430 | return tok; |
431 | } | | 431 | } |
432 | | | 432 | |
433 | /* | | 433 | /* |
434 | * Look up the definition of a name in the symbol table. This symbol must | | 434 | * Look up the definition of a name in the symbol table. This symbol must |
435 | * either be a keyword or a symbol of the type required by symtyp (label, | | 435 | * either be a keyword or a symbol of the type required by symtyp (label, |
436 | * member, tag, ...). | | 436 | * member, tag, ...). |
437 | */ | | 437 | */ |
438 | extern int | | 438 | extern int |
439 | lex_name(const char *yytext, size_t yyleng) | | 439 | lex_name(const char *yytext, size_t yyleng) |
440 | { | | 440 | { |
441 | | | 441 | |
442 | sym_t *sym = symtab_search(yytext); | | 442 | sym_t *sym = symtab_search(yytext); |
443 | if (sym != NULL && sym->s_keyword != NULL) | | 443 | if (sym != NULL && sym->s_keyword != NULL) |
444 | return lex_keyword(sym); | | 444 | return lex_keyword(sym); |
445 | | | 445 | |
446 | sbuf_t *sb = xmalloc(sizeof(*sb)); | | 446 | sbuf_t *sb = xmalloc(sizeof(*sb)); |
447 | sb->sb_len = yyleng; | | 447 | sb->sb_len = yyleng; |
448 | sb->sb_sym = sym; | | 448 | sb->sb_sym = sym; |
449 | yylval.y_name = sb; | | 449 | yylval.y_name = sb; |
450 | | | 450 | |
451 | if (sym != NULL) { | | 451 | if (sym != NULL) { |
452 | lint_assert(block_level >= sym->s_block_level); | | 452 | lint_assert(block_level >= sym->s_block_level); |
453 | sb->sb_name = sym->s_name; | | 453 | sb->sb_name = sym->s_name; |
454 | return sym->s_scl == TYPEDEF ? T_TYPENAME : T_NAME; | | 454 | return sym->s_scl == TYPEDEF ? T_TYPENAME : T_NAME; |
455 | } | | 455 | } |
456 | | | 456 | |
457 | char *name = block_zero_alloc(yyleng + 1); | | 457 | char *name = block_zero_alloc(yyleng + 1); |
458 | (void)memcpy(name, yytext, yyleng + 1); | | 458 | (void)memcpy(name, yytext, yyleng + 1); |
459 | sb->sb_name = name; | | 459 | sb->sb_name = name; |
460 | return T_NAME; | | 460 | return T_NAME; |
461 | } | | 461 | } |
462 | | | 462 | |
463 | int | | 463 | int |
464 | lex_integer_constant(const char *yytext, size_t yyleng, int base) | | 464 | lex_integer_constant(const char *yytext, size_t yyleng, int base) |
465 | { | | 465 | { |
466 | /* C11 6.4.4.1p5 */ | | 466 | /* C11 6.4.4.1p5 */ |
467 | static const tspec_t suffix_type[2][3] = { | | 467 | static const tspec_t suffix_type[2][3] = { |
468 | { INT, LONG, QUAD, }, | | 468 | { INT, LONG, QUAD, }, |
469 | { UINT, ULONG, UQUAD, } | | 469 | { UINT, ULONG, UQUAD, } |
470 | }; | | 470 | }; |
471 | | | 471 | |
472 | const char *cp = yytext; | | 472 | const char *cp = yytext; |
473 | size_t len = yyleng; | | 473 | size_t len = yyleng; |
474 | | | 474 | |
475 | /* skip 0[xX] or 0[bB] */ | | 475 | /* skip 0[xX] or 0[bB] */ |
476 | if (base == 16 || base == 2) { | | 476 | if (base == 16 || base == 2) { |
477 | cp += 2; | | 477 | cp += 2; |
478 | len -= 2; | | 478 | len -= 2; |
479 | } | | 479 | } |
480 | | | 480 | |
481 | /* read suffixes */ | | 481 | /* read suffixes */ |
482 | unsigned l_suffix = 0, u_suffix = 0; | | 482 | unsigned l_suffix = 0, u_suffix = 0; |
483 | for (;; len--) { | | 483 | for (;; len--) { |
484 | char c = cp[len - 1]; | | 484 | char c = cp[len - 1]; |
485 | if (c == 'l' || c == 'L') | | 485 | if (c == 'l' || c == 'L') |
486 | l_suffix++; | | 486 | l_suffix++; |
487 | else if (c == 'u' || c == 'U') | | 487 | else if (c == 'u' || c == 'U') |
488 | u_suffix++; | | 488 | u_suffix++; |
489 | else | | 489 | else |
490 | break; | | 490 | break; |
491 | } | | 491 | } |
492 | if (l_suffix > 2 || u_suffix > 1) { | | 492 | if (l_suffix > 2 || u_suffix > 1) { |
493 | /* malformed integer constant */ | | 493 | /* malformed integer constant */ |
494 | warning(251); | | 494 | warning(251); |
495 | if (l_suffix > 2) | | 495 | if (l_suffix > 2) |
496 | l_suffix = 2; | | 496 | l_suffix = 2; |
497 | if (u_suffix > 1) | | 497 | if (u_suffix > 1) |
498 | u_suffix = 1; | | 498 | u_suffix = 1; |
499 | } | | 499 | } |
500 | if (!allow_c90 && u_suffix > 0) { | | 500 | if (!allow_c90 && u_suffix > 0) { |
501 | /* suffix U is illegal in traditional C */ | | 501 | /* suffix U is illegal in traditional C */ |
502 | warning(97); | | 502 | warning(97); |
503 | } | | 503 | } |
504 | tspec_t typ = suffix_type[u_suffix][l_suffix]; | | 504 | tspec_t typ = suffix_type[u_suffix][l_suffix]; |
505 | | | 505 | |
506 | bool warned = false; | | 506 | bool warned = false; |
507 | errno = 0; | | 507 | errno = 0; |
508 | char *eptr; | | 508 | char *eptr; |
509 | uint64_t uq = (uint64_t)strtoull(cp, &eptr, base); | | 509 | uint64_t uq = (uint64_t)strtoull(cp, &eptr, base); |
510 | lint_assert(eptr == cp + len); | | 510 | lint_assert(eptr == cp + len); |
511 | if (errno != 0) { | | 511 | if (errno != 0) { |
512 | /* integer constant out of range */ | | 512 | /* integer constant out of range */ |
513 | warning(252); | | 513 | warning(252); |
514 | warned = true; | | 514 | warned = true; |
515 | } | | 515 | } |
516 | | | 516 | |
517 | if (any_query_enabled && base == 8 && uq != 0) { | | 517 | if (any_query_enabled && base == 8 && uq != 0) { |
518 | /* octal number '%.*s' */ | | 518 | /* octal number '%.*s' */ |
519 | query_message(8, (int)len, cp); | | 519 | query_message(8, (int)len, cp); |
520 | } | | 520 | } |
521 | | | 521 | |
522 | /* | | 522 | /* |
523 | * If the value is too big for the current type, we must choose | | 523 | * If the value is too big for the current type, we must choose |
524 | * another type. | | 524 | * another type. |
525 | */ | | 525 | */ |
526 | bool ansiu = false; | | 526 | bool ansiu = false; |
527 | switch (typ) { | | 527 | switch (typ) { |
528 | case INT: | | 528 | case INT: |
529 | if (uq <= TARG_INT_MAX) { | | 529 | if (uq <= TARG_INT_MAX) { |
530 | /* ok */ | | 530 | /* ok */ |
531 | } else if (uq <= TARG_UINT_MAX && base != 10) { | | 531 | } else if (uq <= TARG_UINT_MAX && base != 10) { |
532 | typ = UINT; | | 532 | typ = UINT; |
533 | } else if (uq <= TARG_LONG_MAX) { | | 533 | } else if (uq <= TARG_LONG_MAX) { |
534 | typ = LONG; | | 534 | typ = LONG; |
535 | } else { | | 535 | } else { |
536 | typ = ULONG; | | 536 | typ = ULONG; |
537 | if (uq > TARG_ULONG_MAX && !warned) { | | 537 | if (uq > TARG_ULONG_MAX && !warned) { |
538 | /* integer constant out of range */ | | 538 | /* integer constant out of range */ |
539 | warning(252); | | 539 | warning(252); |
540 | } | | 540 | } |
541 | } | | 541 | } |
542 | if (typ == UINT || typ == ULONG) { | | 542 | if (typ == UINT || typ == ULONG) { |
543 | if (!allow_c90) { | | 543 | if (!allow_c90) { |
544 | typ = LONG; | | 544 | typ = LONG; |
545 | } else if (allow_trad) { | | 545 | } else if (allow_trad) { |
546 | /* | | 546 | /* |
547 | * Remember that the constant is unsigned | | 547 | * Remember that the constant is unsigned |
548 | * only in ANSI C. | | 548 | * only in ANSI C. |
549 | */ | | 549 | */ |
550 | ansiu = true; | | 550 | ansiu = true; |
551 | } | | 551 | } |
552 | } | | 552 | } |
553 | break; | | 553 | break; |
554 | case UINT: | | 554 | case UINT: |
555 | if (uq > TARG_UINT_MAX) { | | 555 | if (uq > TARG_UINT_MAX) { |
556 | typ = ULONG; | | 556 | typ = ULONG; |
557 | if (uq > TARG_ULONG_MAX && !warned) { | | 557 | if (uq > TARG_ULONG_MAX && !warned) { |
558 | /* integer constant out of range */ | | 558 | /* integer constant out of range */ |
559 | warning(252); | | 559 | warning(252); |
560 | } | | 560 | } |
561 | } | | 561 | } |
562 | break; | | 562 | break; |
563 | case LONG: | | 563 | case LONG: |
564 | if (uq > TARG_LONG_MAX && allow_c90) { | | 564 | if (uq > TARG_LONG_MAX && allow_c90) { |
565 | typ = ULONG; | | 565 | typ = ULONG; |
566 | if (allow_trad) | | 566 | if (allow_trad) |
567 | ansiu = true; | | 567 | ansiu = true; |
568 | if (uq > TARG_ULONG_MAX && !warned) { | | 568 | if (uq > TARG_ULONG_MAX && !warned) { |
569 | /* integer constant out of range */ | | 569 | /* integer constant out of range */ |
570 | warning(252); | | 570 | warning(252); |
571 | } | | 571 | } |
572 | } | | 572 | } |
573 | break; | | 573 | break; |
574 | case ULONG: | | 574 | case ULONG: |
575 | if (uq > TARG_ULONG_MAX && !warned) { | | 575 | if (uq > TARG_ULONG_MAX && !warned) { |
576 | /* integer constant out of range */ | | 576 | /* integer constant out of range */ |
577 | warning(252); | | 577 | warning(252); |
578 | } | | 578 | } |
579 | break; | | 579 | break; |
580 | case QUAD: | | 580 | case QUAD: |
581 | if (uq > TARG_QUAD_MAX && allow_c90) | | 581 | if (uq > TARG_QUAD_MAX && allow_c90) |
582 | typ = UQUAD; | | 582 | typ = UQUAD; |
583 | break; | | 583 | break; |
584 | case UQUAD: | | 584 | case UQUAD: |
585 | if (uq > TARG_UQUAD_MAX && !warned) { | | 585 | if (uq > TARG_UQUAD_MAX && !warned) { |
586 | /* integer constant out of range */ | | 586 | /* integer constant out of range */ |
587 | warning(252); | | 587 | warning(252); |
588 | } | | 588 | } |
589 | break; | | 589 | break; |
590 | default: | | 590 | default: |
591 | break; | | 591 | break; |
592 | } | | 592 | } |
593 | | | 593 | |
594 | uq = (uint64_t)convert_integer((int64_t)uq, typ, 0); | | 594 | uq = (uint64_t)convert_integer((int64_t)uq, typ, 0); |
595 | | | 595 | |
596 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); | | 596 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); |
597 | yylval.y_val->v_tspec = typ; | | 597 | yylval.y_val->v_tspec = typ; |
598 | yylval.y_val->v_unsigned_since_c90 = ansiu; | | 598 | yylval.y_val->v_unsigned_since_c90 = ansiu; |
599 | yylval.y_val->v_quad = (int64_t)uq; | | 599 | yylval.y_val->v_quad = (int64_t)uq; |
600 | | | 600 | |
601 | return T_CON; | | 601 | return T_CON; |
602 | } | | 602 | } |
603 | | | 603 | |
604 | /* | | 604 | /* |
605 | * Extend or truncate q to match t. If t is signed, sign-extend. | | 605 | * Extend or truncate q to match t. If t is signed, sign-extend. |
606 | * | | 606 | * |
607 | * len is the number of significant bits. If len is 0, len is set | | 607 | * len is the number of significant bits. If len is 0, len is set |
608 | * to the width of type t. | | 608 | * to the width of type t. |
609 | */ | | 609 | */ |
610 | int64_t | | 610 | int64_t |
611 | convert_integer(int64_t q, tspec_t t, unsigned int len) | | 611 | convert_integer(int64_t q, tspec_t t, unsigned int len) |
612 | { | | 612 | { |
613 | | | 613 | |
614 | if (len == 0) | | 614 | if (len == 0) |
615 | len = size_in_bits(t); | | 615 | len = size_in_bits(t); |
616 | | | 616 | |
617 | uint64_t vbits = value_bits(len); | | 617 | uint64_t vbits = value_bits(len); |
618 | return t == PTR || is_uinteger(t) || ((q & bit(len - 1)) == 0) | | 618 | return t == PTR || is_uinteger(t) || ((q & bit(len - 1)) == 0) |
619 | ? (int64_t)(q & vbits) | | 619 | ? (int64_t)(q & vbits) |
620 | : (int64_t)(q | ~vbits); | | 620 | : (int64_t)(q | ~vbits); |
621 | } | | 621 | } |
622 | | | 622 | |
623 | int | | 623 | int |
624 | lex_floating_constant(const char *yytext, size_t yyleng) | | 624 | lex_floating_constant(const char *yytext, size_t yyleng) |
625 | { | | 625 | { |
626 | const char *cp = yytext; | | 626 | const char *cp = yytext; |
627 | size_t len = yyleng; | | 627 | size_t len = yyleng; |
628 | | | 628 | |
629 | if (cp[len - 1] == 'i') | | 629 | if (cp[len - 1] == 'i') |
630 | len--; /* imaginary, do nothing for now */ | | 630 | len--; /* imaginary, do nothing for now */ |
631 | | | 631 | |
632 | char c = cp[len - 1]; | | 632 | char c = cp[len - 1]; |
633 | tspec_t typ; | | 633 | tspec_t typ; |
634 | if (c == 'f' || c == 'F') { | | 634 | if (c == 'f' || c == 'F') { |
635 | typ = FLOAT; | | 635 | typ = FLOAT; |
636 | len--; | | 636 | len--; |
637 | } else if (c == 'l' || c == 'L') { | | 637 | } else if (c == 'l' || c == 'L') { |
638 | typ = LDOUBLE; | | 638 | typ = LDOUBLE; |
639 | len--; | | 639 | len--; |
640 | } else | | 640 | } else |
641 | typ = DOUBLE; | | 641 | typ = DOUBLE; |
642 | | | 642 | |
643 | if (!allow_c90 && typ != DOUBLE) { | | 643 | if (!allow_c90 && typ != DOUBLE) { |
644 | /* suffixes F and L are illegal in traditional C */ | | 644 | /* suffixes F and L are illegal in traditional C */ |
645 | warning(98); | | 645 | warning(98); |
646 | } | | 646 | } |
647 | | | 647 | |
648 | errno = 0; | | 648 | errno = 0; |
649 | char *eptr; | | 649 | char *eptr; |
650 | long double ld = strtold(cp, &eptr); | | 650 | long double ld = strtold(cp, &eptr); |
651 | lint_assert(eptr == cp + len); | | 651 | lint_assert(eptr == cp + len); |
652 | if (errno != 0) | | 652 | if (errno != 0) { |
653 | /* floating-point constant out of range */ | | 653 | /* floating-point constant out of range */ |
654 | warning(248); | | 654 | warning(248); |
655 | | | 655 | } else if (typ == FLOAT) { |
656 | if (typ == FLOAT) { | | | |
657 | ld = (float)ld; | | 656 | ld = (float)ld; |
658 | if (isfinite(ld) == 0) { | | 657 | if (isfinite(ld) == 0) { |
659 | /* floating-point constant out of range */ | | 658 | /* floating-point constant out of range */ |
660 | warning(248); | | 659 | warning(248); |
661 | ld = ld > 0 ? FLT_MAX : -FLT_MAX; | | 660 | ld = ld > 0 ? FLT_MAX : -FLT_MAX; |
662 | } | | 661 | } |
663 | } else if (typ == DOUBLE) { | | 662 | } else if (typ == DOUBLE) { |
664 | ld = (double)ld; | | 663 | ld = (double)ld; |
665 | if (isfinite(ld) == 0) { | | 664 | if (isfinite(ld) == 0) { |
666 | /* floating-point constant out of range */ | | 665 | /* floating-point constant out of range */ |
667 | warning(248); | | 666 | warning(248); |
668 | ld = ld > 0 ? DBL_MAX : -DBL_MAX; | | 667 | ld = ld > 0 ? DBL_MAX : -DBL_MAX; |
669 | } | | 668 | } |
670 | } | | 669 | } |
671 | | | 670 | |
672 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); | | 671 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); |
673 | yylval.y_val->v_tspec = typ; | | 672 | yylval.y_val->v_tspec = typ; |
674 | yylval.y_val->v_ldbl = ld; | | 673 | yylval.y_val->v_ldbl = ld; |
675 | | | 674 | |
676 | return T_CON; | | 675 | return T_CON; |
677 | } | | 676 | } |
678 | | | 677 | |
679 | int | | 678 | int |
680 | lex_operator(int t, op_t o) | | 679 | lex_operator(int t, op_t o) |
681 | { | | 680 | { |
682 | | | 681 | |
683 | yylval.y_op = o; | | 682 | yylval.y_op = o; |
684 | return t; | | 683 | return t; |
685 | } | | 684 | } |
686 | | | 685 | |
687 | static int prev_byte = -1; | | 686 | static int prev_byte = -1; |
688 | | | 687 | |
689 | static int | | 688 | static int |
690 | read_escaped_oct(int c) | | 689 | read_escaped_oct(int c) |
691 | { | | 690 | { |
692 | int n = 3; | | 691 | int n = 3; |
693 | int value = 0; | | 692 | int value = 0; |
694 | do { | | 693 | do { |
695 | value = (value << 3) + (c - '0'); | | 694 | value = (value << 3) + (c - '0'); |
696 | c = read_byte(); | | 695 | c = read_byte(); |
697 | } while (--n > 0 && '0' <= c && c <= '7'); | | 696 | } while (--n > 0 && '0' <= c && c <= '7'); |
698 | prev_byte = c; | | 697 | prev_byte = c; |
699 | if (value > TARG_UCHAR_MAX) { | | 698 | if (value > TARG_UCHAR_MAX) { |
700 | /* character escape does not fit in character */ | | 699 | /* character escape does not fit in character */ |
701 | warning(76); | | 700 | warning(76); |
702 | value &= CHAR_MASK; | | 701 | value &= CHAR_MASK; |
703 | } | | 702 | } |
704 | return value; | | 703 | return value; |
705 | } | | 704 | } |
706 | | | 705 | |
707 | static unsigned int | | 706 | static unsigned int |
708 | read_escaped_hex(int c) | | 707 | read_escaped_hex(int c) |
709 | { | | 708 | { |
710 | if (!allow_c90) | | 709 | if (!allow_c90) |
711 | /* \x undefined in traditional C */ | | 710 | /* \x undefined in traditional C */ |
712 | warning(82); | | 711 | warning(82); |
713 | unsigned int value = 0; | | 712 | unsigned int value = 0; |
714 | int state = 0; /* 0 = no digits, 1 = OK, 2 = overflow */ | | 713 | int state = 0; /* 0 = no digits, 1 = OK, 2 = overflow */ |
715 | while (c = read_byte(), isxdigit(c)) { | | 714 | while (c = read_byte(), isxdigit(c)) { |
716 | c = isdigit(c) ? c - '0' : toupper(c) - 'A' + 10; | | 715 | c = isdigit(c) ? c - '0' : toupper(c) - 'A' + 10; |
717 | value = (value << 4) + c; | | 716 | value = (value << 4) + c; |
718 | if (state == 2) | | 717 | if (state == 2) |
719 | continue; | | 718 | continue; |
720 | if ((value & ~CHAR_MASK) != 0) { | | 719 | if ((value & ~CHAR_MASK) != 0) { |
721 | /* overflow in hex escape */ | | 720 | /* overflow in hex escape */ |
722 | warning(75); | | 721 | warning(75); |
723 | state = 2; | | 722 | state = 2; |
724 | } else { | | 723 | } else { |
725 | state = 1; | | 724 | state = 1; |
726 | } | | 725 | } |
727 | } | | 726 | } |
728 | prev_byte = c; | | 727 | prev_byte = c; |
729 | if (state == 0) { | | 728 | if (state == 0) { |
730 | /* no hex digits follow \x */ | | 729 | /* no hex digits follow \x */ |
731 | error(74); | | 730 | error(74); |
732 | } | | 731 | } |
733 | if (state == 2) | | 732 | if (state == 2) |
734 | value &= CHAR_MASK; | | 733 | value &= CHAR_MASK; |
735 | return value; | | 734 | return value; |
736 | } | | 735 | } |
737 | | | 736 | |
738 | static int | | 737 | static int |
739 | read_escaped_backslash(int delim) | | 738 | read_escaped_backslash(int delim) |
740 | { | | 739 | { |
741 | int c; | | 740 | int c; |
742 | | | 741 | |
743 | switch (c = read_byte()) { | | 742 | switch (c = read_byte()) { |
744 | case '"': | | 743 | case '"': |
745 | if (!allow_c90 && delim == '\'') | | 744 | if (!allow_c90 && delim == '\'') |
746 | /* \" inside character constants undef... */ | | 745 | /* \" inside character constants undef... */ |
747 | warning(262); | | 746 | warning(262); |
748 | return '"'; | | 747 | return '"'; |
749 | case '\'': | | 748 | case '\'': |
750 | return '\''; | | 749 | return '\''; |
751 | case '?': | | 750 | case '?': |
752 | if (!allow_c90) | | 751 | if (!allow_c90) |
753 | /* \? undefined in traditional C */ | | 752 | /* \? undefined in traditional C */ |
754 | warning(263); | | 753 | warning(263); |
755 | return '?'; | | 754 | return '?'; |
756 | case '\\': | | 755 | case '\\': |
757 | return '\\'; | | 756 | return '\\'; |
758 | case 'a': | | 757 | case 'a': |
759 | if (!allow_c90) | | 758 | if (!allow_c90) |
760 | /* \a undefined in traditional C */ | | 759 | /* \a undefined in traditional C */ |
761 | warning(81); | | 760 | warning(81); |
762 | return '\a'; | | 761 | return '\a'; |
763 | case 'b': | | 762 | case 'b': |
764 | return '\b'; | | 763 | return '\b'; |
765 | case 'f': | | 764 | case 'f': |
766 | return '\f'; | | 765 | return '\f'; |
767 | case 'n': | | 766 | case 'n': |
768 | return '\n'; | | 767 | return '\n'; |
769 | case 'r': | | 768 | case 'r': |
770 | return '\r'; | | 769 | return '\r'; |
771 | case 't': | | 770 | case 't': |
772 | return '\t'; | | 771 | return '\t'; |
773 | case 'v': | | 772 | case 'v': |
774 | if (!allow_c90) | | 773 | if (!allow_c90) |
775 | /* \v undefined in traditional C */ | | 774 | /* \v undefined in traditional C */ |
776 | warning(264); | | 775 | warning(264); |
777 | return '\v'; | | 776 | return '\v'; |
778 | case '8': case '9': | | 777 | case '8': case '9': |
779 | /* bad octal digit %c */ | | 778 | /* bad octal digit %c */ |
780 | warning(77, c); | | 779 | warning(77, c); |
781 | /* FALLTHROUGH */ | | 780 | /* FALLTHROUGH */ |
782 | case '0': case '1': case '2': case '3': | | 781 | case '0': case '1': case '2': case '3': |
783 | case '4': case '5': case '6': case '7': | | 782 | case '4': case '5': case '6': case '7': |
784 | return read_escaped_oct(c); | | 783 | return read_escaped_oct(c); |
785 | case 'x': | | 784 | case 'x': |
786 | return (int)read_escaped_hex(c); | | 785 | return (int)read_escaped_hex(c); |
787 | case '\n': | | 786 | case '\n': |
788 | return -3; | | 787 | return -3; |
789 | case EOF: | | 788 | case EOF: |
790 | return -2; | | 789 | return -2; |
791 | default: | | 790 | default: |
792 | if (isprint(c)) { | | 791 | if (isprint(c)) { |
793 | /* dubious escape \%c */ | | 792 | /* dubious escape \%c */ |
794 | warning(79, c); | | 793 | warning(79, c); |
795 | } else { | | 794 | } else { |
796 | /* dubious escape \%o */ | | 795 | /* dubious escape \%o */ |
797 | warning(80, c); | | 796 | warning(80, c); |
798 | } | | 797 | } |
799 | return c; | | 798 | return c; |
800 | } | | 799 | } |
801 | } | | 800 | } |
802 | | | 801 | |
803 | /* | | 802 | /* |
804 | * Read a character which is part of a character constant or of a string | | 803 | * Read a character which is part of a character constant or of a string |
805 | * and handle escapes. | | 804 | * and handle escapes. |
806 | * | | 805 | * |
807 | * 'delim' is '\'' for character constants and '"' for string literals. | | 806 | * 'delim' is '\'' for character constants and '"' for string literals. |
808 | * | | 807 | * |
809 | * Returns -1 if the end of the character constant or string is reached, | | 808 | * Returns -1 if the end of the character constant or string is reached, |
810 | * -2 if the EOF is reached, and the character otherwise. | | 809 | * -2 if the EOF is reached, and the character otherwise. |
811 | */ | | 810 | */ |
812 | static int | | 811 | static int |
813 | get_escaped_char(int delim) | | 812 | get_escaped_char(int delim) |
814 | { | | 813 | { |
815 | | | 814 | |
816 | int c = prev_byte; | | 815 | int c = prev_byte; |
817 | if (c != -1) | | 816 | if (c != -1) |
818 | prev_byte = -1; | | 817 | prev_byte = -1; |
819 | else | | 818 | else |
820 | c = read_byte(); | | 819 | c = read_byte(); |
821 | | | 820 | |
822 | if (c == delim) | | 821 | if (c == delim) |
823 | return -1; | | 822 | return -1; |
824 | switch (c) { | | 823 | switch (c) { |
825 | case '\n': | | 824 | case '\n': |
826 | if (!allow_c90) { | | 825 | if (!allow_c90) { |
827 | /* newline in string or char constant */ | | 826 | /* newline in string or char constant */ |
828 | error(254); | | 827 | error(254); |
829 | return -2; | | 828 | return -2; |
830 | } | | 829 | } |
831 | return c; | | 830 | return c; |
832 | case '\0': | | 831 | case '\0': |
833 | /* syntax error '%s' */ | | 832 | /* syntax error '%s' */ |
834 | error(249, "EOF or null byte in literal"); | | 833 | error(249, "EOF or null byte in literal"); |
835 | return -2; | | 834 | return -2; |
836 | case EOF: | | 835 | case EOF: |
837 | return -2; | | 836 | return -2; |
838 | case '\\': | | 837 | case '\\': |
839 | c = read_escaped_backslash(delim); | | 838 | c = read_escaped_backslash(delim); |
840 | if (c == -3) | | 839 | if (c == -3) |
841 | return get_escaped_char(delim); | | 840 | return get_escaped_char(delim); |
842 | } | | 841 | } |
843 | return c; | | 842 | return c; |
844 | } | | 843 | } |
845 | | | 844 | |
846 | /* Called if lex found a leading "'". */ | | 845 | /* Called if lex found a leading "'". */ |
847 | int | | 846 | int |
848 | lex_character_constant(void) | | 847 | lex_character_constant(void) |
849 | { | | 848 | { |
850 | size_t n; | | 849 | size_t n; |
851 | int val, c; | | 850 | int val, c; |
852 | | | 851 | |
853 | n = 0; | | 852 | n = 0; |
854 | val = 0; | | 853 | val = 0; |
855 | while ((c = get_escaped_char('\'')) >= 0) { | | 854 | while ((c = get_escaped_char('\'')) >= 0) { |
856 | val = (int)((unsigned int)val << CHAR_SIZE) + c; | | 855 | val = (int)((unsigned int)val << CHAR_SIZE) + c; |
857 | n++; | | 856 | n++; |
858 | } | | 857 | } |
859 | if (c == -2) { | | 858 | if (c == -2) { |
860 | /* unterminated character constant */ | | 859 | /* unterminated character constant */ |
861 | error(253); | | 860 | error(253); |
862 | } else if (n > sizeof(int) || (n > 1 && (pflag || hflag))) { | | 861 | } else if (n > sizeof(int) || (n > 1 && (pflag || hflag))) { |
863 | /* | | 862 | /* |
864 | * XXX: ^^ should rather be sizeof(TARG_INT). Luckily, | | 863 | * XXX: ^^ should rather be sizeof(TARG_INT). Luckily, |
865 | * sizeof(int) is the same on all supported platforms. | | 864 | * sizeof(int) is the same on all supported platforms. |
866 | */ | | 865 | */ |
867 | /* too many characters in character constant */ | | 866 | /* too many characters in character constant */ |
868 | error(71); | | 867 | error(71); |
869 | } else if (n > 1) { | | 868 | } else if (n > 1) { |
870 | /* multi-character character constant */ | | 869 | /* multi-character character constant */ |
871 | warning(294); | | 870 | warning(294); |
872 | } else if (n == 0) { | | 871 | } else if (n == 0) { |
873 | /* empty character constant */ | | 872 | /* empty character constant */ |
874 | error(73); | | 873 | error(73); |
875 | } | | 874 | } |
876 | if (n == 1) | | 875 | if (n == 1) |
877 | val = (int)convert_integer(val, CHAR, CHAR_SIZE); | | 876 | val = (int)convert_integer(val, CHAR, CHAR_SIZE); |
878 | | | 877 | |
879 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); | | 878 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); |
880 | yylval.y_val->v_tspec = INT; | | 879 | yylval.y_val->v_tspec = INT; |
881 | yylval.y_val->v_quad = val; | | 880 | yylval.y_val->v_quad = val; |
882 | | | 881 | |
883 | return T_CON; | | 882 | return T_CON; |
884 | } | | 883 | } |
885 | | | 884 | |
886 | /* | | 885 | /* |
887 | * Called if lex found a leading L\' | | 886 | * Called if lex found a leading L\' |
888 | */ | | 887 | */ |
889 | int | | 888 | int |
890 | lex_wide_character_constant(void) | | 889 | lex_wide_character_constant(void) |
891 | { | | 890 | { |
892 | static char buf[MB_LEN_MAX + 1]; | | 891 | static char buf[MB_LEN_MAX + 1]; |
893 | size_t n, nmax; | | 892 | size_t n, nmax; |
894 | int c; | | 893 | int c; |
895 | wchar_t wc; | | 894 | wchar_t wc; |
896 | | | 895 | |
897 | nmax = MB_CUR_MAX; | | 896 | nmax = MB_CUR_MAX; |
898 | | | 897 | |
899 | n = 0; | | 898 | n = 0; |
900 | while ((c = get_escaped_char('\'')) >= 0) { | | 899 | while ((c = get_escaped_char('\'')) >= 0) { |
901 | if (n < nmax) | | 900 | if (n < nmax) |
902 | buf[n] = (char)c; | | 901 | buf[n] = (char)c; |
903 | n++; | | 902 | n++; |
904 | } | | 903 | } |
905 | | | 904 | |
906 | wc = 0; | | 905 | wc = 0; |
907 | | | 906 | |
908 | if (c == -2) { | | 907 | if (c == -2) { |
909 | /* unterminated character constant */ | | 908 | /* unterminated character constant */ |
910 | error(253); | | 909 | error(253); |
911 | } else if (n == 0) { | | 910 | } else if (n == 0) { |
912 | /* empty character constant */ | | 911 | /* empty character constant */ |
913 | error(73); | | 912 | error(73); |
914 | } else if (n > nmax) { | | 913 | } else if (n > nmax) { |
915 | n = nmax; | | 914 | n = nmax; |
916 | /* too many characters in character constant */ | | 915 | /* too many characters in character constant */ |
917 | error(71); | | 916 | error(71); |
918 | } else { | | 917 | } else { |
919 | buf[n] = '\0'; | | 918 | buf[n] = '\0'; |
920 | (void)mbtowc(NULL, NULL, 0); | | 919 | (void)mbtowc(NULL, NULL, 0); |
921 | if (mbtowc(&wc, buf, nmax) < 0) | | 920 | if (mbtowc(&wc, buf, nmax) < 0) |
922 | /* invalid multibyte character */ | | 921 | /* invalid multibyte character */ |
923 | error(291); | | 922 | error(291); |
924 | } | | 923 | } |
925 | | | 924 | |
926 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); | | 925 | yylval.y_val = xcalloc(1, sizeof(*yylval.y_val)); |
927 | yylval.y_val->v_tspec = WCHAR; | | 926 | yylval.y_val->v_tspec = WCHAR; |
928 | yylval.y_val->v_quad = wc; | | 927 | yylval.y_val->v_quad = wc; |
929 | | | 928 | |
930 | return T_CON; | | 929 | return T_CON; |
931 | } | | 930 | } |
932 | | | 931 | |
933 | /* See https://gcc.gnu.org/onlinedocs/cpp/Preprocessor-Output.html */ | | 932 | /* See https://gcc.gnu.org/onlinedocs/cpp/Preprocessor-Output.html */ |
934 | static void | | 933 | static void |
935 | parse_line_directive_flags(const char *p, | | 934 | parse_line_directive_flags(const char *p, |
936 | bool *is_begin, bool *is_end, bool *is_system) | | 935 | bool *is_begin, bool *is_end, bool *is_system) |
937 | { | | 936 | { |
938 | | | 937 | |
939 | *is_begin = false; | | 938 | *is_begin = false; |
940 | *is_end = false; | | 939 | *is_end = false; |
941 | *is_system = false; | | 940 | *is_system = false; |
942 | | | 941 | |
943 | while (*p != '\0') { | | 942 | while (*p != '\0') { |
944 | const char *word_start, *word_end; | | 943 | const char *word_start, *word_end; |
945 | | | 944 | |
946 | while (ch_isspace(*p)) | | 945 | while (ch_isspace(*p)) |
947 | p++; | | 946 | p++; |
948 | | | 947 | |
949 | word_start = p; | | 948 | word_start = p; |
950 | while (*p != '\0' && !ch_isspace(*p)) | | 949 | while (*p != '\0' && !ch_isspace(*p)) |
951 | p++; | | 950 | p++; |
952 | word_end = p; | | 951 | word_end = p; |
953 | | | 952 | |
954 | if (word_end - word_start == 1 && word_start[0] == '1') | | 953 | if (word_end - word_start == 1 && word_start[0] == '1') |
955 | *is_begin = true; | | 954 | *is_begin = true; |
956 | if (word_end - word_start == 1 && word_start[0] == '2') | | 955 | if (word_end - word_start == 1 && word_start[0] == '2') |
957 | *is_end = true; | | 956 | *is_end = true; |
958 | if (word_end - word_start == 1 && word_start[0] == '3') | | 957 | if (word_end - word_start == 1 && word_start[0] == '3') |
959 | *is_system = true; | | 958 | *is_system = true; |
960 | /* Flag '4' is only interesting for C++. */ | | 959 | /* Flag '4' is only interesting for C++. */ |
961 | } | | 960 | } |
962 | } | | 961 | } |
963 | | | 962 | |
964 | /* | | 963 | /* |
965 | * Called for preprocessor directives. Currently implemented are: | | 964 | * Called for preprocessor directives. Currently implemented are: |
966 | * # pragma [argument...] | | 965 | * # pragma [argument...] |
967 | * # lineno | | 966 | * # lineno |
968 | * # lineno "filename" | | 967 | * # lineno "filename" |
969 | * # lineno "filename" GCC-flag... | | 968 | * # lineno "filename" GCC-flag... |
970 | */ | | 969 | */ |
971 | void | | 970 | void |
972 | lex_directive(const char *yytext) | | 971 | lex_directive(const char *yytext) |
973 | { | | 972 | { |
974 | const char *cp, *fn; | | 973 | const char *cp, *fn; |
975 | char c, *eptr; | | 974 | char c, *eptr; |
976 | size_t fnl; | | 975 | size_t fnl; |
977 | long ln; | | 976 | long ln; |
978 | bool is_begin, is_end, is_system; | | 977 | bool is_begin, is_end, is_system; |
979 | | | 978 | |
980 | static bool first = true; | | 979 | static bool first = true; |
981 | | | 980 | |
982 | /* Go to first non-whitespace after # */ | | 981 | /* Go to first non-whitespace after # */ |
983 | for (cp = yytext + 1; (c = *cp) == ' ' || c == '\t'; cp++) | | 982 | for (cp = yytext + 1; (c = *cp) == ' ' || c == '\t'; cp++) |
984 | continue; | | 983 | continue; |
985 | | | 984 | |
986 | if (!ch_isdigit(c)) { | | 985 | if (!ch_isdigit(c)) { |
987 | if (strncmp(cp, "pragma", 6) == 0 && ch_isspace(cp[6])) | | 986 | if (strncmp(cp, "pragma", 6) == 0 && ch_isspace(cp[6])) |
988 | return; | | 987 | return; |
989 | error: | | 988 | error: |
990 | /* undefined or invalid # directive */ | | 989 | /* undefined or invalid # directive */ |
991 | warning(255); | | 990 | warning(255); |
992 | return; | | 991 | return; |
993 | } | | 992 | } |
994 | ln = strtol(--cp, &eptr, 10); | | 993 | ln = strtol(--cp, &eptr, 10); |
995 | if (eptr == cp) | | 994 | if (eptr == cp) |
996 | goto error; | | 995 | goto error; |
997 | if ((c = *(cp = eptr)) != ' ' && c != '\t' && c != '\0') | | 996 | if ((c = *(cp = eptr)) != ' ' && c != '\t' && c != '\0') |
998 | goto error; | | 997 | goto error; |
999 | while ((c = *cp++) == ' ' || c == '\t') | | 998 | while ((c = *cp++) == ' ' || c == '\t') |
1000 | continue; | | 999 | continue; |
1001 | if (c != '\0') { | | 1000 | if (c != '\0') { |
1002 | if (c != '"') | | 1001 | if (c != '"') |
1003 | goto error; | | 1002 | goto error; |
1004 | fn = cp; | | 1003 | fn = cp; |
1005 | while ((c = *cp) != '"' && c != '\0') | | 1004 | while ((c = *cp) != '"' && c != '\0') |
1006 | cp++; | | 1005 | cp++; |
1007 | if (c != '"') | | 1006 | if (c != '"') |
1008 | goto error; | | 1007 | goto error; |
1009 | if ((fnl = cp++ - fn) > PATH_MAX) | | 1008 | if ((fnl = cp++ - fn) > PATH_MAX) |
1010 | goto error; | | 1009 | goto error; |
1011 | /* empty string means stdin */ | | 1010 | /* empty string means stdin */ |
1012 | if (fnl == 0) { | | 1011 | if (fnl == 0) { |
1013 | fn = "{standard input}"; | | 1012 | fn = "{standard input}"; |
1014 | fnl = 16; /* strlen (fn) */ | | 1013 | fnl = 16; /* strlen (fn) */ |
1015 | } | | 1014 | } |
1016 | curr_pos.p_file = record_filename(fn, fnl); | | 1015 | curr_pos.p_file = record_filename(fn, fnl); |
1017 | /* | | 1016 | /* |
1018 | * If this is the first directive, the name is the name | | 1017 | * If this is the first directive, the name is the name |
1019 | * of the C source file as specified at the command line. | | 1018 | * of the C source file as specified at the command line. |
1020 | * It is written to the output file. | | 1019 | * It is written to the output file. |
1021 | */ | | 1020 | */ |
1022 | if (first) { | | 1021 | if (first) { |
1023 | csrc_pos.p_file = curr_pos.p_file; | | 1022 | csrc_pos.p_file = curr_pos.p_file; |
1024 | outsrc(transform_filename(curr_pos.p_file, | | 1023 | outsrc(transform_filename(curr_pos.p_file, |
1025 | strlen(curr_pos.p_file))); | | 1024 | strlen(curr_pos.p_file))); |
1026 | first = false; | | 1025 | first = false; |
1027 | } | | 1026 | } |
1028 | | | 1027 | |
1029 | parse_line_directive_flags(cp, &is_begin, &is_end, &is_system); | | 1028 | parse_line_directive_flags(cp, &is_begin, &is_end, &is_system); |
1030 | update_location(curr_pos.p_file, (int)ln, is_begin, is_end); | | 1029 | update_location(curr_pos.p_file, (int)ln, is_begin, is_end); |
1031 | in_system_header = is_system; | | 1030 | in_system_header = is_system; |
1032 | } | | 1031 | } |
1033 | curr_pos.p_line = (int)ln - 1; | | 1032 | curr_pos.p_line = (int)ln - 1; |
1034 | curr_pos.p_uniq = 0; | | 1033 | curr_pos.p_uniq = 0; |
1035 | if (curr_pos.p_file == csrc_pos.p_file) { | | 1034 | if (curr_pos.p_file == csrc_pos.p_file) { |
1036 | csrc_pos.p_line = (int)ln - 1; | | 1035 | csrc_pos.p_line = (int)ln - 1; |
1037 | csrc_pos.p_uniq = 0; | | 1036 | csrc_pos.p_uniq = 0; |
1038 | } | | 1037 | } |
1039 | } | | 1038 | } |
1040 | | | 1039 | |
1041 | /* | | 1040 | /* |
1042 | * Handle lint comments such as ARGSUSED. | | 1041 | * Handle lint comments such as ARGSUSED. |
1043 | * | | 1042 | * |
1044 | * If one of these comments is recognized, the argument, if any, is | | 1043 | * If one of these comments is recognized, the argument, if any, is |
1045 | * parsed and a function which handles this comment is called. | | 1044 | * parsed and a function which handles this comment is called. |
1046 | */ | | 1045 | */ |
1047 | void | | 1046 | void |
1048 | lex_comment(void) | | 1047 | lex_comment(void) |
1049 | { | | 1048 | { |
1050 | int c; | | 1049 | int c; |
1051 | static const struct { | | 1050 | static const struct { |
1052 | const char *keywd; | | 1051 | const char *keywd; |
1053 | bool arg; | | 1052 | bool arg; |
1054 | void (*func)(int); | | 1053 | void (*func)(int); |
1055 | } keywtab[] = { | | 1054 | } keywtab[] = { |
1056 | { "ARGSUSED", true, argsused }, | | 1055 | { "ARGSUSED", true, argsused }, |
1057 | { "BITFIELDTYPE", false, bitfieldtype }, | | 1056 | { "BITFIELDTYPE", false, bitfieldtype }, |
1058 | { "CONSTCOND", false, constcond }, | | 1057 | { "CONSTCOND", false, constcond }, |
1059 | { "CONSTANTCOND", false, constcond }, | | 1058 | { "CONSTANTCOND", false, constcond }, |
1060 | { "CONSTANTCONDITION", false, constcond }, | | 1059 | { "CONSTANTCONDITION", false, constcond }, |
1061 | { "FALLTHRU", false, fallthru }, | | 1060 | { "FALLTHRU", false, fallthru }, |
1062 | { "FALLTHROUGH", false, fallthru }, | | 1061 | { "FALLTHROUGH", false, fallthru }, |
1063 | { "FALL THROUGH", false, fallthru }, | | 1062 | { "FALL THROUGH", false, fallthru }, |
1064 | { "fallthrough", false, fallthru }, | | 1063 | { "fallthrough", false, fallthru }, |
1065 | { "LINTLIBRARY", false, lintlib }, | | 1064 | { "LINTLIBRARY", false, lintlib }, |
1066 | { "LINTED", true, linted }, | | 1065 | { "LINTED", true, linted }, |
1067 | { "LONGLONG", false, longlong }, | | 1066 | { "LONGLONG", false, longlong }, |
1068 | { "NOSTRICT", true, linted }, | | 1067 | { "NOSTRICT", true, linted }, |
1069 | { "NOTREACHED", false, not_reached }, | | 1068 | { "NOTREACHED", false, not_reached }, |
1070 | { "PRINTFLIKE", true, printflike }, | | 1069 | { "PRINTFLIKE", true, printflike }, |
1071 | { "PROTOLIB", true, protolib }, | | 1070 | { "PROTOLIB", true, protolib }, |
1072 | { "SCANFLIKE", true, scanflike }, | | 1071 | { "SCANFLIKE", true, scanflike }, |
1073 | { "VARARGS", true, varargs }, | | 1072 | { "VARARGS", true, varargs }, |
1074 | }; | | 1073 | }; |
1075 | char keywd[32]; | | 1074 | char keywd[32]; |
1076 | char arg[32]; | | 1075 | char arg[32]; |
1077 | size_t l, i; | | 1076 | size_t l, i; |
1078 | int a; | | 1077 | int a; |
1079 | | | 1078 | |
1080 | bool seen_end_of_comment = false; | | 1079 | bool seen_end_of_comment = false; |
1081 | | | 1080 | |
1082 | /* Skip whitespace after the start of the comment */ | | 1081 | /* Skip whitespace after the start of the comment */ |
1083 | while (c = read_byte(), isspace(c)) | | 1082 | while (c = read_byte(), isspace(c)) |
1084 | continue; | | 1083 | continue; |
1085 | | | 1084 | |
1086 | /* Read the potential keyword to keywd */ | | 1085 | /* Read the potential keyword to keywd */ |
1087 | l = 0; | | 1086 | l = 0; |
1088 | while (c != EOF && l < sizeof(keywd) - 1 && | | 1087 | while (c != EOF && l < sizeof(keywd) - 1 && |
1089 | (isalpha(c) || isspace(c))) { | | 1088 | (isalpha(c) || isspace(c))) { |
1090 | if (islower(c) && l > 0 && ch_isupper(keywd[0])) | | 1089 | if (islower(c) && l > 0 && ch_isupper(keywd[0])) |
1091 | break; | | 1090 | break; |
1092 | keywd[l++] = (char)c; | | 1091 | keywd[l++] = (char)c; |
1093 | c = read_byte(); | | 1092 | c = read_byte(); |
1094 | } | | 1093 | } |
1095 | while (l > 0 && ch_isspace(keywd[l - 1])) | | 1094 | while (l > 0 && ch_isspace(keywd[l - 1])) |
1096 | l--; | | 1095 | l--; |
1097 | keywd[l] = '\0'; | | 1096 | keywd[l] = '\0'; |
1098 | | | 1097 | |
1099 | /* look for the keyword */ | | 1098 | /* look for the keyword */ |
1100 | for (i = 0; i < sizeof(keywtab) / sizeof(keywtab[0]); i++) { | | 1099 | for (i = 0; i < sizeof(keywtab) / sizeof(keywtab[0]); i++) { |
1101 | if (strcmp(keywtab[i].keywd, keywd) == 0) | | 1100 | if (strcmp(keywtab[i].keywd, keywd) == 0) |
1102 | break; | | 1101 | break; |
1103 | } | | 1102 | } |
1104 | if (i == sizeof(keywtab) / sizeof(keywtab[0])) | | 1103 | if (i == sizeof(keywtab) / sizeof(keywtab[0])) |
1105 | goto skip_rest; | | 1104 | goto skip_rest; |
1106 | | | 1105 | |
1107 | /* skip whitespace after the keyword */ | | 1106 | /* skip whitespace after the keyword */ |
1108 | while (isspace(c)) | | 1107 | while (isspace(c)) |
1109 | c = read_byte(); | | 1108 | c = read_byte(); |
1110 | | | 1109 | |
1111 | /* read the argument, if the keyword accepts one and there is one */ | | 1110 | /* read the argument, if the keyword accepts one and there is one */ |
1112 | l = 0; | | 1111 | l = 0; |
1113 | if (keywtab[i].arg) { | | 1112 | if (keywtab[i].arg) { |
1114 | while (isdigit(c) && l < sizeof(arg) - 1) { | | 1113 | while (isdigit(c) && l < sizeof(arg) - 1) { |
1115 | arg[l++] = (char)c; | | 1114 | arg[l++] = (char)c; |
1116 | c = read_byte(); | | 1115 | c = read_byte(); |
1117 | } | | 1116 | } |
1118 | } | | 1117 | } |
1119 | arg[l] = '\0'; | | 1118 | arg[l] = '\0'; |
1120 | a = l != 0 ? atoi(arg) : -1; | | 1119 | a = l != 0 ? atoi(arg) : -1; |
1121 | | | 1120 | |
1122 | /* skip whitespace after the argument */ | | 1121 | /* skip whitespace after the argument */ |
1123 | while (isspace(c)) | | 1122 | while (isspace(c)) |
1124 | c = read_byte(); | | 1123 | c = read_byte(); |
1125 | | | 1124 | |
1126 | seen_end_of_comment = c == '*' && (c = read_byte()) == '/'; | | 1125 | seen_end_of_comment = c == '*' && (c = read_byte()) == '/'; |
1127 | if (!seen_end_of_comment && keywtab[i].func != linted) | | 1126 | if (!seen_end_of_comment && keywtab[i].func != linted) |
1128 | /* extra characters in lint comment */ | | 1127 | /* extra characters in lint comment */ |
1129 | warning(257); | | 1128 | warning(257); |
1130 | | | 1129 | |
1131 | if (keywtab[i].func != NULL) | | 1130 | if (keywtab[i].func != NULL) |
1132 | keywtab[i].func(a); | | 1131 | keywtab[i].func(a); |
1133 | | | 1132 | |
1134 | skip_rest: | | 1133 | skip_rest: |
1135 | while (!seen_end_of_comment) { | | 1134 | while (!seen_end_of_comment) { |
1136 | int lc = c; | | 1135 | int lc = c; |
1137 | if ((c = read_byte()) == EOF) { | | 1136 | if ((c = read_byte()) == EOF) { |
1138 | /* unterminated comment */ | | 1137 | /* unterminated comment */ |
1139 | error(256); | | 1138 | error(256); |
1140 | break; | | 1139 | break; |
1141 | } | | 1140 | } |
1142 | if (lc == '*' && c == '/') | | 1141 | if (lc == '*' && c == '/') |
1143 | seen_end_of_comment = true; | | 1142 | seen_end_of_comment = true; |
1144 | } | | 1143 | } |
1145 | } | | 1144 | } |
1146 | | | 1145 | |
1147 | void | | 1146 | void |
1148 | lex_slash_slash_comment(void) | | 1147 | lex_slash_slash_comment(void) |
1149 | { | | 1148 | { |
1150 | int c; | | 1149 | int c; |
1151 | | | 1150 | |
1152 | if (!allow_c99 && !allow_gcc) | | 1151 | if (!allow_c99 && !allow_gcc) |
1153 | /* %s does not support // comments */ | | 1152 | /* %s does not support // comments */ |
1154 | gnuism(312, allow_c90 ? "C90" : "traditional C"); | | 1153 | gnuism(312, allow_c90 ? "C90" : "traditional C"); |
1155 | | | 1154 | |
1156 | while ((c = read_byte()) != EOF && c != '\n') | | 1155 | while ((c = read_byte()) != EOF && c != '\n') |
1157 | continue; | | 1156 | continue; |
1158 | } | | 1157 | } |
1159 | | | 1158 | |
1160 | /* | | 1159 | /* |
1161 | * Clear flags for lint comments LINTED, LONGLONG and CONSTCOND. | | 1160 | * Clear flags for lint comments LINTED, LONGLONG and CONSTCOND. |
1162 | * clear_warn_flags is called after function definitions and global and | | 1161 | * clear_warn_flags is called after function definitions and global and |
1163 | * local declarations and definitions. It is also called between | | 1162 | * local declarations and definitions. It is also called between |
1164 | * the controlling expression and the body of control statements | | 1163 | * the controlling expression and the body of control statements |
1165 | * (if, switch, for, while). | | 1164 | * (if, switch, for, while). |
1166 | */ | | 1165 | */ |
1167 | void | | 1166 | void |
1168 | clear_warn_flags(void) | | 1167 | clear_warn_flags(void) |
1169 | { | | 1168 | { |
1170 | | | 1169 | |
1171 | lwarn = LWARN_ALL; | | 1170 | lwarn = LWARN_ALL; |
1172 | quadflg = false; | | 1171 | quadflg = false; |
1173 | constcond_flag = false; | | 1172 | constcond_flag = false; |
1174 | } | | 1173 | } |
1175 | | | 1174 | |
1176 | int | | 1175 | int |
1177 | lex_string(void) | | 1176 | lex_string(void) |
1178 | { | | 1177 | { |
1179 | unsigned char *s; | | 1178 | unsigned char *s; |
1180 | int c; | | 1179 | int c; |
1181 | size_t len, max; | | 1180 | size_t len, max; |
1182 | | | 1181 | |
1183 | s = xmalloc(max = 64); | | 1182 | s = xmalloc(max = 64); |
1184 | | | 1183 | |
1185 | len = 0; | | 1184 | len = 0; |
1186 | while ((c = get_escaped_char('"')) >= 0) { | | 1185 | while ((c = get_escaped_char('"')) >= 0) { |
1187 | /* +1 to reserve space for a trailing NUL character */ | | 1186 | /* +1 to reserve space for a trailing NUL character */ |
1188 | if (len + 1 == max) | | 1187 | if (len + 1 == max) |
1189 | s = xrealloc(s, max *= 2); | | 1188 | s = xrealloc(s, max *= 2); |
1190 | s[len++] = (char)c; | | 1189 | s[len++] = (char)c; |
1191 | } | | 1190 | } |
1192 | s[len] = '\0'; | | 1191 | s[len] = '\0'; |
1193 | if (c == -2) | | 1192 | if (c == -2) |
1194 | /* unterminated string constant */ | | 1193 | /* unterminated string constant */ |
1195 | error(258); | | 1194 | error(258); |
1196 | | | 1195 | |
1197 | strg_t *strg = xcalloc(1, sizeof(*strg)); | | 1196 | strg_t *strg = xcalloc(1, sizeof(*strg)); |
1198 | strg->st_char = true; | | 1197 | strg->st_char = true; |
1199 | strg->st_len = len; | | 1198 | strg->st_len = len; |
1200 | strg->st_mem = s; | | 1199 | strg->st_mem = s; |
1201 | | | 1200 | |
1202 | yylval.y_string = strg; | | 1201 | yylval.y_string = strg; |
1203 | return T_STRING; | | 1202 | return T_STRING; |
1204 | } | | 1203 | } |
1205 | | | 1204 | |
1206 | int | | 1205 | int |
1207 | lex_wide_string(void) | | 1206 | lex_wide_string(void) |
1208 | { | | 1207 | { |
1209 | int c, n; | | 1208 | int c, n; |
1210 | | | 1209 | |
1211 | size_t len = 0, max = 64; | | 1210 | size_t len = 0, max = 64; |
1212 | char *s = xmalloc(max); | | 1211 | char *s = xmalloc(max); |
1213 | while ((c = get_escaped_char('"')) >= 0) { | | 1212 | while ((c = get_escaped_char('"')) >= 0) { |
1214 | /* +1 to save space for a trailing NUL character */ | | 1213 | /* +1 to save space for a trailing NUL character */ |
1215 | if (len + 1 >= max) | | 1214 | if (len + 1 >= max) |
1216 | s = xrealloc(s, max *= 2); | | 1215 | s = xrealloc(s, max *= 2); |
1217 | s[len++] = (char)c; | | 1216 | s[len++] = (char)c; |
1218 | } | | 1217 | } |
1219 | s[len] = '\0'; | | 1218 | s[len] = '\0'; |
1220 | if (c == -2) | | 1219 | if (c == -2) |
1221 | /* unterminated string constant */ | | 1220 | /* unterminated string constant */ |
1222 | error(258); | | 1221 | error(258); |
1223 | | | 1222 | |
1224 | /* get length of wide-character string */ | | 1223 | /* get length of wide-character string */ |
1225 | (void)mblen(NULL, 0); | | 1224 | (void)mblen(NULL, 0); |
1226 | size_t wlen = 0; | | 1225 | size_t wlen = 0; |
1227 | for (size_t i = 0; i < len; i += n, wlen++) { | | 1226 | for (size_t i = 0; i < len; i += n, wlen++) { |
1228 | if ((n = mblen(&s[i], MB_CUR_MAX)) == -1) { | | 1227 | if ((n = mblen(&s[i], MB_CUR_MAX)) == -1) { |
1229 | /* invalid multibyte character */ | | 1228 | /* invalid multibyte character */ |
1230 | error(291); | | 1229 | error(291); |
1231 | break; | | 1230 | break; |
1232 | } | | 1231 | } |
1233 | if (n == 0) | | 1232 | if (n == 0) |
1234 | n = 1; | | 1233 | n = 1; |
1235 | } | | 1234 | } |
1236 | | | 1235 | |
1237 | wchar_t *ws = xmalloc((wlen + 1) * sizeof(*ws)); | | 1236 | wchar_t *ws = xmalloc((wlen + 1) * sizeof(*ws)); |
1238 | size_t wi = 0; | | 1237 | size_t wi = 0; |
1239 | /* convert from multibyte to wide char */ | | 1238 | /* convert from multibyte to wide char */ |
1240 | (void)mbtowc(NULL, NULL, 0); | | 1239 | (void)mbtowc(NULL, NULL, 0); |
1241 | for (size_t i = 0; i < len; i += n, wi++) { | | 1240 | for (size_t i = 0; i < len; i += n, wi++) { |
1242 | if ((n = mbtowc(&ws[wi], &s[i], MB_CUR_MAX)) == -1) | | 1241 | if ((n = mbtowc(&ws[wi], &s[i], MB_CUR_MAX)) == -1) |
1243 | break; | | 1242 | break; |
1244 | if (n == 0) | | 1243 | if (n == 0) |
1245 | n = 1; | | 1244 | n = 1; |
1246 | } | | 1245 | } |
1247 | ws[wi] = 0; | | 1246 | ws[wi] = 0; |
1248 | free(s); | | 1247 | free(s); |
1249 | | | 1248 | |
1250 | strg_t *strg = xcalloc(1, sizeof(*strg)); | | 1249 | strg_t *strg = xcalloc(1, sizeof(*strg)); |
1251 | strg->st_char = false; | | 1250 | strg->st_char = false; |
1252 | strg->st_len = wlen; | | 1251 | strg->st_len = wlen; |
1253 | strg->st_mem = ws; | | 1252 | strg->st_mem = ws; |
1254 | | | 1253 | |
1255 | yylval.y_string = strg; | | 1254 | yylval.y_string = strg; |
1256 | return T_STRING; | | 1255 | return T_STRING; |
1257 | } | | 1256 | } |
1258 | | | 1257 | |
1259 | void | | 1258 | void |
1260 | lex_next_line(void) | | 1259 | lex_next_line(void) |
1261 | { | | 1260 | { |
1262 | curr_pos.p_line++; | | 1261 | curr_pos.p_line++; |
1263 | curr_pos.p_uniq = 0; | | 1262 | curr_pos.p_uniq = 0; |
1264 | debug_step("parsing %s:%d", curr_pos.p_file, curr_pos.p_line); | | 1263 | debug_step("parsing %s:%d", curr_pos.p_file, curr_pos.p_line); |
1265 | if (curr_pos.p_file == csrc_pos.p_file) { | | 1264 | if (curr_pos.p_file == csrc_pos.p_file) { |
1266 | csrc_pos.p_line++; | | 1265 | csrc_pos.p_line++; |
1267 | csrc_pos.p_uniq = 0; | | 1266 | csrc_pos.p_uniq = 0; |
1268 | } | | 1267 | } |
1269 | } | | 1268 | } |
1270 | | | 1269 | |
1271 | void | | 1270 | void |
1272 | lex_unknown_character(int c) | | 1271 | lex_unknown_character(int c) |
1273 | { | | 1272 | { |
1274 | | | 1273 | |
1275 | /* unknown character \%o */ | | 1274 | /* unknown character \%o */ |
1276 | error(250, c); | | 1275 | error(250, c); |
1277 | } | | 1276 | } |
1278 | | | 1277 | |
1279 | /* | | 1278 | /* |
1280 | * The scanner does not create new symbol table entries for symbols it cannot | | 1279 | * The scanner does not create new symbol table entries for symbols it cannot |
1281 | * find in the symbol table. This is to avoid putting undeclared symbols into | | 1280 | * find in the symbol table. This is to avoid putting undeclared symbols into |
1282 | * the symbol table if a syntax error occurs. | | 1281 | * the symbol table if a syntax error occurs. |
1283 | * | | 1282 | * |
1284 | * getsym is called as soon as it is probably ok to put the symbol in the | | 1283 | * getsym is called as soon as it is probably ok to put the symbol in the |
1285 | * symbol table. It is still possible that symbols are put in the symbol | | 1284 | * symbol table. It is still possible that symbols are put in the symbol |
1286 | * table that are not completely declared due to syntax errors. To avoid too | | 1285 | * table that are not completely declared due to syntax errors. To avoid too |
1287 | * many problems in this case, symbols get type 'int' in getsym. | | 1286 | * many problems in this case, symbols get type 'int' in getsym. |
1288 | * | | 1287 | * |
1289 | * XXX calls to getsym should be delayed until declare_1_* is called. | | 1288 | * XXX calls to getsym should be delayed until declare_1_* is called. |
1290 | */ | | 1289 | */ |
1291 | sym_t * | | 1290 | sym_t * |
1292 | getsym(sbuf_t *sb) | | 1291 | getsym(sbuf_t *sb) |
1293 | { | | 1292 | { |
1294 | | | 1293 | |
1295 | sym_t *sym = sb->sb_sym; | | 1294 | sym_t *sym = sb->sb_sym; |
1296 | | | 1295 | |
1297 | /* | | 1296 | /* |
1298 | * During member declaration it is possible that name() looked | | 1297 | * During member declaration it is possible that name() looked |
1299 | * for symbols of type FVFT, although it should have looked for | | 1298 | * for symbols of type FVFT, although it should have looked for |
1300 | * symbols of type FTAG. Same can happen for labels. Both cases | | 1299 | * symbols of type FTAG. Same can happen for labels. Both cases |
1301 | * are compensated here. | | 1300 | * are compensated here. |
1302 | */ | | 1301 | */ |
1303 | if (symtyp == FMEMBER || symtyp == FLABEL) { | | 1302 | if (symtyp == FMEMBER || symtyp == FLABEL) { |
1304 | if (sym == NULL || sym->s_kind == FVFT) | | 1303 | if (sym == NULL || sym->s_kind == FVFT) |
1305 | sym = symtab_search(sb->sb_name); | | 1304 | sym = symtab_search(sb->sb_name); |
1306 | } | | 1305 | } |
1307 | | | 1306 | |
1308 | if (sym != NULL) { | | 1307 | if (sym != NULL) { |
1309 | lint_assert(sym->s_kind == symtyp); | | 1308 | lint_assert(sym->s_kind == symtyp); |
1310 | symtyp = FVFT; | | 1309 | symtyp = FVFT; |
1311 | free(sb); | | 1310 | free(sb); |
1312 | return sym; | | 1311 | return sym; |
1313 | } | | 1312 | } |
1314 | | | 1313 | |
1315 | /* create a new symbol table entry */ | | 1314 | /* create a new symbol table entry */ |
1316 | | | 1315 | |
1317 | /* labels must always be allocated at level 1 (outermost block) */ | | 1316 | /* labels must always be allocated at level 1 (outermost block) */ |
1318 | dinfo_t *di; | | 1317 | dinfo_t *di; |
1319 | if (symtyp == FLABEL) { | | 1318 | if (symtyp == FLABEL) { |
1320 | sym = level_zero_alloc(1, sizeof(*sym)); | | 1319 | sym = level_zero_alloc(1, sizeof(*sym)); |
1321 | char *s = level_zero_alloc(1, sb->sb_len + 1); | | 1320 | char *s = level_zero_alloc(1, sb->sb_len + 1); |
1322 | (void)memcpy(s, sb->sb_name, sb->sb_len + 1); | | 1321 | (void)memcpy(s, sb->sb_name, sb->sb_len + 1); |
1323 | sym->s_name = s; | | 1322 | sym->s_name = s; |
1324 | sym->s_block_level = 1; | | 1323 | sym->s_block_level = 1; |
1325 | di = dcs; | | 1324 | di = dcs; |
1326 | while (di->d_enclosing != NULL && | | 1325 | while (di->d_enclosing != NULL && |
1327 | di->d_enclosing->d_enclosing != NULL) | | 1326 | di->d_enclosing->d_enclosing != NULL) |
1328 | di = di->d_enclosing; | | 1327 | di = di->d_enclosing; |
1329 | lint_assert(di->d_kind == DK_AUTO); | | 1328 | lint_assert(di->d_kind == DK_AUTO); |
1330 | } else { | | 1329 | } else { |
1331 | sym = block_zero_alloc(sizeof(*sym)); | | 1330 | sym = block_zero_alloc(sizeof(*sym)); |
1332 | sym->s_name = sb->sb_name; | | 1331 | sym->s_name = sb->sb_name; |
1333 | sym->s_block_level = block_level; | | 1332 | sym->s_block_level = block_level; |
1334 | di = dcs; | | 1333 | di = dcs; |
1335 | } | | 1334 | } |
1336 | | | 1335 | |
1337 | UNIQUE_CURR_POS(sym->s_def_pos); | | 1336 | UNIQUE_CURR_POS(sym->s_def_pos); |
1338 | if ((sym->s_kind = symtyp) != FLABEL) | | 1337 | if ((sym->s_kind = symtyp) != FLABEL) |
1339 | sym->s_type = gettyp(INT); | | 1338 | sym->s_type = gettyp(INT); |
1340 | | | 1339 | |
1341 | symtyp = FVFT; | | 1340 | symtyp = FVFT; |
1342 | | | 1341 | |
1343 | if (!in_gcc_attribute) { | | 1342 | if (!in_gcc_attribute) { |
1344 | symtab_add(sym); | | 1343 | symtab_add(sym); |
1345 | | | 1344 | |
1346 | *di->d_ldlsym = sym; | | 1345 | *di->d_ldlsym = sym; |
1347 | di->d_ldlsym = &sym->s_level_next; | | 1346 | di->d_ldlsym = &sym->s_level_next; |
1348 | } | | 1347 | } |
1349 | | | 1348 | |
1350 | free(sb); | | 1349 | free(sb); |
1351 | return sym; | | 1350 | return sym; |
1352 | } | | 1351 | } |
1353 | | | 1352 | |
1354 | /* | | 1353 | /* |
1355 | * Construct a temporary symbol. The symbol name starts with a digit to avoid | | 1354 | * Construct a temporary symbol. The symbol name starts with a digit to avoid |
1356 | * name clashes with other identifiers. | | 1355 | * name clashes with other identifiers. |
1357 | */ | | 1356 | */ |
1358 | sym_t * | | 1357 | sym_t * |
1359 | mktempsym(type_t *tp) | | 1358 | mktempsym(type_t *tp) |
1360 | { | | 1359 | { |
1361 | static unsigned n = 0; | | 1360 | static unsigned n = 0; |
1362 | char *s = level_zero_alloc((size_t)block_level, 64); | | 1361 | char *s = level_zero_alloc((size_t)block_level, 64); |
1363 | sym_t *sym = block_zero_alloc(sizeof(*sym)); | | 1362 | sym_t *sym = block_zero_alloc(sizeof(*sym)); |
1364 | scl_t scl; | | 1363 | scl_t scl; |
1365 | | | 1364 | |
1366 | (void)snprintf(s, 64, "%.8u_tmp", n++); | | 1365 | (void)snprintf(s, 64, "%.8u_tmp", n++); |
1367 | | | 1366 | |
1368 | scl = dcs->d_scl; | | 1367 | scl = dcs->d_scl; |
1369 | if (scl == NOSCL) | | 1368 | if (scl == NOSCL) |
1370 | scl = block_level > 0 ? AUTO : EXTERN; | | 1369 | scl = block_level > 0 ? AUTO : EXTERN; |
1371 | | | 1370 | |
1372 | sym->s_name = s; | | 1371 | sym->s_name = s; |
1373 | sym->s_type = tp; | | 1372 | sym->s_type = tp; |
1374 | sym->s_block_level = block_level; | | 1373 | sym->s_block_level = block_level; |
1375 | sym->s_scl = scl; | | 1374 | sym->s_scl = scl; |
1376 | sym->s_kind = FVFT; | | 1375 | sym->s_kind = FVFT; |
1377 | sym->s_used = true; | | 1376 | sym->s_used = true; |
1378 | sym->s_set = true; | | 1377 | sym->s_set = true; |
1379 | | | 1378 | |
1380 | symtab_add(sym); | | 1379 | symtab_add(sym); |
1381 | | | 1380 | |
1382 | *dcs->d_ldlsym = sym; | | 1381 | *dcs->d_ldlsym = sym; |
1383 | dcs->d_ldlsym = &sym->s_level_next; | | 1382 | dcs->d_ldlsym = &sym->s_level_next; |
1384 | | | 1383 | |
1385 | return sym; | | 1384 | return sym; |
1386 | } | | 1385 | } |
1387 | | | 1386 | |
1388 | /* Remove a symbol forever from the symbol table. */ | | 1387 | /* Remove a symbol forever from the symbol table. */ |
1389 | void | | 1388 | void |
1390 | rmsym(sym_t *sym) | | 1389 | rmsym(sym_t *sym) |
1391 | { | | 1390 | { |
1392 | | | 1391 | |
1393 | debug_step("rmsym '%s' %s '%s'", | | 1392 | debug_step("rmsym '%s' %s '%s'", |
1394 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); | | 1393 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); |
1395 | symtab_remove(sym); | | 1394 | symtab_remove(sym); |
1396 | | | 1395 | |
1397 | /* avoid that the symbol will later be put back to the symbol table */ | | 1396 | /* avoid that the symbol will later be put back to the symbol table */ |
1398 | sym->s_block_level = -1; | | 1397 | sym->s_block_level = -1; |
1399 | } | | 1398 | } |
1400 | | | 1399 | |
1401 | /* | | 1400 | /* |
1402 | * Remove all symbols from the symbol table that have the same level as the | | 1401 | * Remove all symbols from the symbol table that have the same level as the |
1403 | * given symbol. | | 1402 | * given symbol. |
1404 | */ | | 1403 | */ |
1405 | void | | 1404 | void |
1406 | rmsyms(sym_t *syms) | | 1405 | rmsyms(sym_t *syms) |
1407 | { | | 1406 | { |
1408 | sym_t *sym; | | 1407 | sym_t *sym; |
1409 | | | 1408 | |
1410 | /* Note the use of s_level_next instead of s_symtab_next. */ | | 1409 | /* Note the use of s_level_next instead of s_symtab_next. */ |
1411 | for (sym = syms; sym != NULL; sym = sym->s_level_next) { | | 1410 | for (sym = syms; sym != NULL; sym = sym->s_level_next) { |
1412 | if (sym->s_block_level != -1) { | | 1411 | if (sym->s_block_level != -1) { |
1413 | debug_step("rmsyms '%s' %s '%s'", | | 1412 | debug_step("rmsyms '%s' %s '%s'", |
1414 | sym->s_name, symt_name(sym->s_kind), | | 1413 | sym->s_name, symt_name(sym->s_kind), |
1415 | type_name(sym->s_type)); | | 1414 | type_name(sym->s_type)); |
1416 | symtab_remove(sym); | | 1415 | symtab_remove(sym); |
1417 | sym->s_symtab_ref = NULL; | | 1416 | sym->s_symtab_ref = NULL; |
1418 | } | | 1417 | } |
1419 | } | | 1418 | } |
1420 | } | | 1419 | } |
1421 | | | 1420 | |
1422 | /* Put a symbol into the symbol table. */ | | 1421 | /* Put a symbol into the symbol table. */ |
1423 | void | | 1422 | void |
1424 | inssym(int level, sym_t *sym) | | 1423 | inssym(int level, sym_t *sym) |
1425 | { | | 1424 | { |
1426 | | | 1425 | |
1427 | debug_step("inssym '%s' %s '%s'", | | 1426 | debug_step("inssym '%s' %s '%s'", |
1428 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); | | 1427 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); |
1429 | symtab_add(sym); | | 1428 | symtab_add(sym); |
1430 | sym->s_block_level = level; | | 1429 | sym->s_block_level = level; |
1431 | | | 1430 | |
1432 | /* | | 1431 | /* |
1433 | * Placing the inner symbols to the beginning of the list ensures | | 1432 | * Placing the inner symbols to the beginning of the list ensures |
1434 | * that these symbols are preferred over symbols from the outer | | 1433 | * that these symbols are preferred over symbols from the outer |
1435 | * blocks that happen to have the same name. | | 1434 | * blocks that happen to have the same name. |
1436 | */ | | 1435 | */ |
1437 | const sym_t *next = sym->s_symtab_next; | | 1436 | const sym_t *next = sym->s_symtab_next; |
1438 | if (next != NULL) | | 1437 | if (next != NULL) |
1439 | lint_assert(sym->s_block_level >= next->s_block_level); | | 1438 | lint_assert(sym->s_block_level >= next->s_block_level); |
1440 | } | | 1439 | } |
1441 | | | 1440 | |
1442 | /* Called at level 0 after syntax errors. */ | | 1441 | /* Called at level 0 after syntax errors. */ |
1443 | void | | 1442 | void |
1444 | clean_up_after_error(void) | | 1443 | clean_up_after_error(void) |
1445 | { | | 1444 | { |
1446 | | | 1445 | |
1447 | symtab_remove_locals(); | | 1446 | symtab_remove_locals(); |
1448 | | | 1447 | |
1449 | while (mem_block_level > 0) | | 1448 | while (mem_block_level > 0) |
1450 | level_free_all(mem_block_level--); | | 1449 | level_free_all(mem_block_level--); |
1451 | } | | 1450 | } |
1452 | | | 1451 | |
1453 | /* Create a new symbol with the same name as an existing symbol. */ | | 1452 | /* Create a new symbol with the same name as an existing symbol. */ |
1454 | sym_t * | | 1453 | sym_t * |
1455 | pushdown(const sym_t *sym) | | 1454 | pushdown(const sym_t *sym) |
1456 | { | | 1455 | { |
1457 | sym_t *nsym; | | 1456 | sym_t *nsym; |
1458 | | | 1457 | |
1459 | debug_step("pushdown '%s' %s '%s'", | | 1458 | debug_step("pushdown '%s' %s '%s'", |
1460 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); | | 1459 | sym->s_name, symt_name(sym->s_kind), type_name(sym->s_type)); |
1461 | nsym = block_zero_alloc(sizeof(*nsym)); | | 1460 | nsym = block_zero_alloc(sizeof(*nsym)); |
1462 | lint_assert(sym->s_block_level <= block_level); | | 1461 | lint_assert(sym->s_block_level <= block_level); |
1463 | nsym->s_name = sym->s_name; | | 1462 | nsym->s_name = sym->s_name; |
1464 | UNIQUE_CURR_POS(nsym->s_def_pos); | | 1463 | UNIQUE_CURR_POS(nsym->s_def_pos); |
1465 | nsym->s_kind = sym->s_kind; | | 1464 | nsym->s_kind = sym->s_kind; |
1466 | nsym->s_block_level = block_level; | | 1465 | nsym->s_block_level = block_level; |
1467 | | | 1466 | |
1468 | symtab_add(nsym); | | 1467 | symtab_add(nsym); |
1469 | | | 1468 | |
1470 | *dcs->d_ldlsym = nsym; | | 1469 | *dcs->d_ldlsym = nsym; |
1471 | dcs->d_ldlsym = &nsym->s_level_next; | | 1470 | dcs->d_ldlsym = &nsym->s_level_next; |
1472 | | | 1471 | |
1473 | return nsym; | | 1472 | return nsym; |
1474 | } | | 1473 | } |
1475 | | | 1474 | |
1476 | /* | | 1475 | /* |
1477 | * Free any dynamically allocated memory referenced by | | 1476 | * Free any dynamically allocated memory referenced by |
1478 | * the value stack or yylval. | | 1477 | * the value stack or yylval. |
1479 | * The type of information in yylval is described by tok. | | 1478 | * The type of information in yylval is described by tok. |
1480 | */ | | 1479 | */ |
1481 | void | | 1480 | void |
1482 | freeyyv(void *sp, int tok) | | 1481 | freeyyv(void *sp, int tok) |
1483 | { | | 1482 | { |
1484 | if (tok == T_NAME || tok == T_TYPENAME) { | | 1483 | if (tok == T_NAME || tok == T_TYPENAME) { |
1485 | sbuf_t *sb = *(sbuf_t **)sp; | | 1484 | sbuf_t *sb = *(sbuf_t **)sp; |
1486 | free(sb); | | 1485 | free(sb); |
1487 | } else if (tok == T_CON) { | | 1486 | } else if (tok == T_CON) { |
1488 | val_t *val = *(val_t **)sp; | | 1487 | val_t *val = *(val_t **)sp; |
1489 | free(val); | | 1488 | free(val); |
1490 | } else if (tok == T_STRING) { | | 1489 | } else if (tok == T_STRING) { |
1491 | strg_t *strg = *(strg_t **)sp; | | 1490 | strg_t *strg = *(strg_t **)sp; |
1492 | free(strg->st_mem); | | 1491 | free(strg->st_mem); |
1493 | free(strg); | | 1492 | free(strg); |
1494 | } | | 1493 | } |
1495 | } | | 1494 | } |