X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;ds=sidebyside;f=lexer.c;h=82bccd05a960748a17a9fdc78f94bb8bcfe34b18;hb=975933a9f57835385489ecdeda04a65894705eca;hp=bfc8875a4bd118e85a40d78c0dd3a2996c6ee8d7;hpb=45a23dceb12ed1d9c0813061800855edf19cf8f9;p=cparser diff --git a/lexer.c b/lexer.c index bfc8875..82bccd0 100644 --- a/lexer.c +++ b/lexer.c @@ -7,6 +7,7 @@ #include "adt/strset.h" #include "adt/util.h" #include "type_t.h" +#include "target_architecture.h" #include #include @@ -22,6 +23,14 @@ #define strtold(s, e) strtod(s, e) #endif +#if defined HAS_SIGNED_CHAR +typedef signed char char_type; +#elif defined HAS_UNSIGNED_CHAR +typedef unsigned char char_type; +#else +# error signedness of char not determined +#endif + static int c; token_t lexer_token; symbol_t *symbol_L; @@ -80,7 +89,7 @@ static inline void put_back(int pc) //assert(bufpos < buf+MAX_PUTBACK || *bufpos == pc); char *p = buf + (bufpos - buf); - *p = pc; + *p = (char) pc; /* going backwards in the buffer is legal as long as it's not more often * than MAX_PUTBACK */ @@ -239,14 +248,14 @@ static void parse_symbol(void) symbol_t *symbol; char *string; - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); while(1) { switch(c) { DIGITS SYMBOL_CHARS - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); break; @@ -269,39 +278,78 @@ end_symbol: } } -static void parse_integer_suffix(void) +static void parse_integer_suffix(bool is_oct_hex) { + bool is_unsigned = false; + bool min_long = false; + bool min_longlong = false; + if(c == 'U' || c == 'u') { + is_unsigned = true; next_char(); if(c == 'L' || c == 'l') { + min_long = true; next_char(); if(c == 'L' || c == 'l') { + min_longlong = true; next_char(); - lexer_token.datatype = type_ulonglong; - } else { - lexer_token.datatype = type_ulong; } - } else { - lexer_token.datatype = type_uint; } } else if(c == 'l' || c == 'L') { + min_long = true; next_char(); if(c == 'l' || c == 'L') { + min_longlong = true; next_char(); if(c == 'u' || c == 'U') { + is_unsigned = true; next_char(); - lexer_token.datatype = type_ulonglong; - } else { - lexer_token.datatype = type_longlong; } } else if(c == 'u' || c == 'U') { + is_unsigned = true; next_char(); lexer_token.datatype = type_ulong; - } else { - lexer_token.datatype = type_int; } + } + + if(!is_unsigned) { + long long v = lexer_token.v.intvalue; + if(!min_long) { + if(v >= TARGET_INT_MIN && v <= TARGET_INT_MAX) { + lexer_token.datatype = type_int; + return; + } else if(is_oct_hex && v >= 0 && v <= TARGET_UINT_MAX) { + lexer_token.datatype = type_uint; + return; + } + } + if(!min_longlong) { + if(v >= TARGET_LONG_MIN && v <= TARGET_LONG_MAX) { + lexer_token.datatype = type_long; + return; + } else if(is_oct_hex && v >= 0 && v <= TARGET_ULONG_MAX) { + lexer_token.datatype = type_ulong; + return; + } + } + unsigned long long uv = (unsigned long long) v; + if(is_oct_hex && uv > (unsigned long long) TARGET_LONGLONG_MAX) { + lexer_token.datatype = type_ulonglong; + return; + } + + lexer_token.datatype = type_longlong; } else { - lexer_token.datatype = type_int; + unsigned long long v = (unsigned long long) lexer_token.v.intvalue; + if(!min_long && v <= TARGET_UINT_MAX) { + lexer_token.datatype = type_uint; + return; + } + if(!min_longlong && v <= TARGET_ULONG_MAX) { + lexer_token.datatype = type_ulong; + return; + } + lexer_token.datatype = type_ulonglong; } } @@ -336,7 +384,7 @@ static unsigned long long parse_int_string(const char *s, const char **endptr, i case 16: for (;; ++s) { /* check for overrun */ - if (v <= 0x1000000000000000ULL) + if (v >= 0x1000000000000000ULL) break; switch (tolower(*s)) { case '0': v <<= 4; break; @@ -356,14 +404,14 @@ static unsigned long long parse_int_string(const char *s, const char **endptr, i case 'e': v <<= 4; v |= 0xe; break; case 'f': v <<= 4; v |= 0xf; break; default: - break; + goto end; } } break; case 8: for (;; ++s) { /* check for overrun */ - if (v <= 0x2000000000000000ULL) + if (v >= 0x2000000000000000ULL) break; switch (tolower(*s)) { case '0': v <<= 3; break; @@ -375,7 +423,7 @@ static unsigned long long parse_int_string(const char *s, const char **endptr, i case '6': v <<= 3; v |= 6; break; case '7': v <<= 3; v |= 7; break; default: - break; + goto end; } } break; @@ -396,7 +444,7 @@ static unsigned long long parse_int_string(const char *s, const char **endptr, i case '8': v *= 10; v += 8; break; case '9': v *= 10; v += 9; break; default: - break; + goto end; } } break; @@ -404,6 +452,7 @@ static unsigned long long parse_int_string(const char *s, const char **endptr, i assert(0); break; } +end: *endptr = s; return v; } @@ -414,7 +463,7 @@ static void parse_number_hex(void) next_char(); while(isxdigit(c)) { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } obstack_1grow(&symbol_obstack, '\0'); @@ -437,7 +486,7 @@ static void parse_number_hex(void) } obstack_free(&symbol_obstack, string); - parse_integer_suffix(); + parse_integer_suffix(true); } static inline bool is_octal_digit(int chr) @@ -448,7 +497,7 @@ static inline bool is_octal_digit(int chr) static void parse_number_oct(void) { while(is_octal_digit(c)) { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } obstack_1grow(&symbol_obstack, '\0'); @@ -462,14 +511,14 @@ static void parse_number_oct(void) } obstack_free(&symbol_obstack, string); - parse_integer_suffix(); + parse_integer_suffix(true); } static void parse_number_dec(void) { bool is_float = false; while(isdigit(c)) { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } @@ -478,7 +527,7 @@ static void parse_number_dec(void) next_char(); while(isdigit(c)) { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } is_float = true; @@ -488,12 +537,12 @@ static void parse_number_dec(void) next_char(); if(c == '-' || c == '+') { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } while(isdigit(c)) { - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); } is_float = true; @@ -502,8 +551,8 @@ static void parse_number_dec(void) obstack_1grow(&symbol_obstack, '\0'); char *string = obstack_finish(&symbol_obstack); - const char *endptr; if(is_float) { + char *endptr; lexer_token.type = T_FLOATINGPOINT; lexer_token.v.floatvalue = strtold(string, &endptr); @@ -513,6 +562,7 @@ static void parse_number_dec(void) parse_floating_suffix(); } else { + const char *endptr; lexer_token.type = T_INTEGER; lexer_token.v.intvalue = parse_int_string(string, &endptr, 10); @@ -520,7 +570,7 @@ static void parse_number_dec(void) parse_error("invalid number literal"); } - parse_integer_suffix(); + parse_integer_suffix(false); } obstack_free(&symbol_obstack, string); } @@ -573,7 +623,7 @@ static int parse_octal_sequence(const int first_digit) if (!is_octal_digit(c)) return value; value = 8 * value + c - '0'; next_char(); - return value; + return (char_type)value; } static int parse_hex_sequence(void) @@ -592,7 +642,7 @@ static int parse_hex_sequence(void) next_char(); } - return value; + return (char_type)value; } static int parse_escape_sequence(void) @@ -665,7 +715,7 @@ static void parse_string_literal(void) switch(c) { case '\\': tc = parse_escape_sequence(); - obstack_1grow(&symbol_obstack, tc); + obstack_1grow(&symbol_obstack, (char) tc); break; case EOF: @@ -680,7 +730,7 @@ static void parse_string_literal(void) goto end_of_string; default: - obstack_1grow(&symbol_obstack, c); + obstack_1grow(&symbol_obstack, (char) c); next_char(); break; } @@ -704,6 +754,66 @@ end_of_string: lexer_token.v.string = result; } +static void parse_wide_string_literal(void) +{ + const unsigned start_linenr = lexer_token.source_position.linenr; + + assert(c == '"'); + next_char(); + + while(1) { + switch(c) { + case '\\': { + wchar_rep_t tc = parse_escape_sequence(); + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + break; + } + + case EOF: + error_prefix_at(lexer_token.source_position.input_name, + start_linenr); + fprintf(stderr, "string has no end\n"); + lexer_token.type = T_ERROR; + return; + + case '"': + next_char(); + goto end_of_string; + + default: { + wchar_rep_t tc = c; + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + next_char(); + break; + } + } + } + +end_of_string:; + + /* TODO: concatenate multiple strings separated by whitespace... */ + + /* add finishing 0 to the string */ + wchar_rep_t nul = L'\0'; + obstack_grow(&symbol_obstack, &nul, sizeof(nul)); + const size_t size = (size_t)obstack_object_size(&symbol_obstack) / sizeof(wchar_rep_t); + const wchar_rep_t *const string = obstack_finish(&symbol_obstack); + +#if 0 /* TODO hash */ + /* check if there is already a copy of the string */ + const wchar_rep_t *const result = strset_insert(&stringset, string); + if(result != string) { + obstack_free(&symbol_obstack, string); + } +#else + const wchar_rep_t *const result = string; +#endif + + lexer_token.type = T_WIDE_STRING_LITERAL; + lexer_token.v.wide_string.begin = result; + lexer_token.v.wide_string.size = size; +} + static void parse_character_constant(void) { eat('\''); @@ -953,8 +1063,7 @@ void lexer_next_preprocessing_token(void) /* might be a wide string ( L"string" ) */ if(c == '"' && (lexer_token.type == T_IDENTIFIER && lexer_token.v.symbol == symbol_L)) { - parse_string_literal(); - return; + parse_wide_string_literal(); } return; @@ -1127,14 +1236,14 @@ void init_lexer(void) { strset_init(&stringset); - type_int = make_atomic_type(ATOMIC_TYPE_INT, TYPE_QUALIFIER_CONST); - type_uint = make_atomic_type(ATOMIC_TYPE_UINT, TYPE_QUALIFIER_CONST); - type_long = make_atomic_type(ATOMIC_TYPE_LONG, TYPE_QUALIFIER_CONST); - type_ulong = make_atomic_type(ATOMIC_TYPE_ULONG, TYPE_QUALIFIER_CONST); + type_int = make_atomic_type(ATOMIC_TYPE_INT, TYPE_QUALIFIER_NONE); + type_uint = make_atomic_type(ATOMIC_TYPE_UINT, TYPE_QUALIFIER_NONE); + type_long = make_atomic_type(ATOMIC_TYPE_LONG, TYPE_QUALIFIER_NONE); + type_ulong = make_atomic_type(ATOMIC_TYPE_ULONG, TYPE_QUALIFIER_NONE); type_longlong = make_atomic_type(ATOMIC_TYPE_LONGLONG, - TYPE_QUALIFIER_CONST); + TYPE_QUALIFIER_NONE); type_ulonglong = make_atomic_type(ATOMIC_TYPE_ULONGLONG, - TYPE_QUALIFIER_CONST); + TYPE_QUALIFIER_NONE); type_float = make_atomic_type(ATOMIC_TYPE_FLOAT, TYPE_QUALIFIER_CONST); type_double = make_atomic_type(ATOMIC_TYPE_DOUBLE, @@ -1164,7 +1273,7 @@ void exit_lexer(void) static __attribute__((unused)) void dbg_pos(const source_position_t source_position) { - fprintf(stdout, "%s:%d\n", source_position.input_name, + fprintf(stdout, "%s:%u\n", source_position.input_name, source_position.linenr); fflush(stdout); }