X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;f=lexer.c;h=1300b3d26f23cda1c90093658115266e03a941ab;hb=93b256acc783d281b7eb29184591a76c8e83e62a;hp=715c851f73ebdf623c69be8afc0b6779f56a2946;hpb=24b92962c733223b5022f0111885796539c760ef;p=cparser diff --git a/lexer.c b/lexer.c index 715c851..1300b3d 100644 --- a/lexer.c +++ b/lexer.c @@ -1,9 +1,11 @@ #include -#include "lexer_t.h" +#include "lexer.h" #include "token_t.h" #include "symbol_table_t.h" #include "adt/error.h" +#include "adt/strset.h" +#include "adt/util.h" #include #include @@ -11,376 +13,588 @@ #include //#define DEBUG_CHARS -#define MAX_PUTBACK 2 - -static -void error_prefix_at(lexer_t *this, const char *input_name, unsigned linenr) +#define MAX_PUTBACK 3 + +static int c; +token_t lexer_token; +symbol_t *symbol_L; +static FILE *input; +static char buf[1024 + MAX_PUTBACK]; +static const char *bufend; +static const char *bufpos; +static strset_t stringset; + +static void error_prefix_at(const char *input_name, unsigned linenr) { - (void) this; - fprintf(stderr, "%s:%d: Error: ", input_name, linenr); + fprintf(stderr, "%s:%u: Error: ", input_name, linenr); } -static -void error_prefix(lexer_t *this) +static void error_prefix(void) { - error_prefix_at(this, this->source_position.input_name, - this->source_position.linenr); + error_prefix_at(lexer_token.source_position.input_name, + lexer_token.source_position.linenr); } -static -void parse_error(lexer_t *this, const char *msg) +static void parse_error(const char *msg) { - error_prefix(this); + error_prefix(); fprintf(stderr, "%s\n", msg); } -static inline -void next_char(lexer_t *this) +static inline void next_real_char(void) { - this->bufpos++; - if(this->bufpos >= this->bufend) { - size_t s = fread(this->buf + MAX_PUTBACK, 1, - sizeof(this->buf) - MAX_PUTBACK, this->input); + bufpos++; + if(bufpos >= bufend) { + size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK, + input); if(s == 0) { - this->c = EOF; + c = EOF; return; } - this->bufpos = this->buf + MAX_PUTBACK; - this->bufend = this->buf + MAX_PUTBACK + s; + bufpos = buf + MAX_PUTBACK; + bufend = buf + MAX_PUTBACK + s; } - this->c = *(this->bufpos); -#ifdef DEBUG_CHARS - printf("nchar '%c'\n", this->c); -#endif + c = *(bufpos); } -static inline -void put_back(lexer_t *this, int c) +static inline void put_back(int pc) { - char *p = (char*) this->bufpos - 1; - this->bufpos--; - assert(p >= this->buf); - *p = c; + assert(bufpos >= buf); + assert(bufpos < buf+MAX_PUTBACK || *bufpos == pc); + + char *p = buf + (bufpos - buf); + *p = pc; + + /* going backwards in the buffer is legal as long as it's not more often + * than MAX_PUTBACK */ + bufpos--; #ifdef DEBUG_CHARS - printf("putback '%c'\n", c); + printf("putback '%c'\n", pc); #endif } -static -int replace_trigraph(lexer_t *this) -{ -#define MATCH_TRIGRAPH(ch,replacement) \ - case ch: \ - this->c = replacement; \ - return 1; - - switch(this->c) { - MATCH_TRIGRAPH('=', '#') - MATCH_TRIGRAPH('(', '[') - MATCH_TRIGRAPH('/', '\\') - MATCH_TRIGRAPH(')', ']') - MATCH_TRIGRAPH('\'', '^') - MATCH_TRIGRAPH('<', '{') - MATCH_TRIGRAPH('!', '|') - MATCH_TRIGRAPH('>', '}') - MATCH_TRIGRAPH('-', '~') +static inline void next_char(void); + +#define MATCH_NEWLINE(code) \ + case '\r': \ + next_char(); \ + if(c == '\n') { \ + next_char(); \ + } \ + lexer_token.source_position.linenr++; \ + code; \ + case '\n': \ + next_char(); \ + lexer_token.source_position.linenr++; \ + code; + +#define eat(c_type) do { assert(c == c_type); next_char(); } while(0) + +static void maybe_concat_lines(void) +{ + eat('\\'); + + switch(c) { + MATCH_NEWLINE(return;) + default: break; } - return 0; -} - -#define SKIP_TRIGRAPHS(no_trigraph_code) \ - case '?': \ - next_char(this); \ - if(this->c != '?') { \ - put_back(this, this->c); \ - this->c = '?'; \ - no_trigraph_code; \ - } \ - next_char(this); \ - if(replace_trigraph(this)) \ - break; \ - put_back(this, '?'); \ - put_back(this, this->c); \ - this->c = '?'; \ - no_trigraph_code; - -#define EAT_NEWLINE(newline_code) \ - if(this->c == '\r') { \ - next_char(this); \ - if(this->c == '\n') \ - next_char(this); \ - this->source_position.linenr++; \ - newline_code; \ - } else if(this->c == '\n') { \ - next_char(this); \ - this->source_position.linenr++; \ - newline_code; \ - } - -static -void parse_symbol(lexer_t *this, token_t *token) + put_back(c); + c = '\\'; +} + +static inline void next_char(void) +{ + next_real_char(); + +#if 0 + /* filter trigraphs */ + if(UNLIKELY(c == '\\')) { + maybe_concat_lines(); + goto end_of_next_char; + } + + if(LIKELY(c != '?')) + goto end_of_next_char; + + next_real_char(); + if(LIKELY(c != '?')) { + put_back(c); + c = '?'; + goto end_of_next_char; + } + + next_real_char(); + switch(c) { + case '=': c = '#'; break; + case '(': c = '['; break; + case '/': c = '\\'; maybe_concat_lines(); break; + case ')': c = ']'; break; + case '\'': c = '^'; break; + case '<': c = '{'; break; + case '!': c = '|'; break; + case '>': c = '}'; break; + case '-': c = '~'; break; + default: + put_back('?'); + put_back(c); + c = '?'; + break; + } + +end_of_next_char: +#endif + (void) maybe_concat_lines; +#ifdef DEBUG_CHARS + printf("nchar '%c'\n", c); +#else + ; +#endif +} + +#define SYMBOL_CHARS \ + case 'a': \ + case 'b': \ + case 'c': \ + case 'd': \ + case 'e': \ + case 'f': \ + case 'g': \ + case 'h': \ + case 'i': \ + case 'j': \ + case 'k': \ + case 'l': \ + case 'm': \ + case 'n': \ + case 'o': \ + case 'p': \ + case 'q': \ + case 'r': \ + case 's': \ + case 't': \ + case 'u': \ + case 'v': \ + case 'w': \ + case 'x': \ + case 'y': \ + case 'z': \ + case 'A': \ + case 'B': \ + case 'C': \ + case 'D': \ + case 'E': \ + case 'F': \ + case 'G': \ + case 'H': \ + case 'I': \ + case 'J': \ + case 'K': \ + case 'L': \ + case 'M': \ + case 'N': \ + case 'O': \ + case 'P': \ + case 'Q': \ + case 'R': \ + case 'S': \ + case 'T': \ + case 'U': \ + case 'V': \ + case 'W': \ + case 'X': \ + case 'Y': \ + case 'Z': \ + case '_': + +#define DIGITS \ + case '0': \ + case '1': \ + case '2': \ + case '3': \ + case '4': \ + case '5': \ + case '6': \ + case '7': \ + case '8': \ + case '9': + +static void parse_symbol(void) { symbol_t *symbol; char *string; - obstack_1grow(&symbol_obstack, this->c); - next_char(this); + obstack_1grow(&symbol_obstack, c); + next_char(); while(1) { - switch(this->c) { - case '\\': - next_char(this); - EAT_NEWLINE(break;) - goto end_symbol; - - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - obstack_1grow(&symbol_obstack, this->c); - next_char(this); + switch(c) { + DIGITS + SYMBOL_CHARS + obstack_1grow(&symbol_obstack, c); + next_char(); break; - case '?': - next_char(this); - if(this->c != '?') { - put_back(this, this->c); - this->c = '?'; - goto end_symbol; - } - next_char(this); - if(replace_trigraph(this)) - break; - put_back(this, '?'); - put_back(this, this->c); - this->c = '?'; - goto end_symbol; - default: goto end_symbol; } } + end_symbol: obstack_1grow(&symbol_obstack, '\0'); string = obstack_finish(&symbol_obstack); symbol = symbol_table_insert(string); - if(symbol->ID > 0) { - token->type = symbol->ID; - } else { - token->type = T_IDENTIFIER; - } - token->v.symbol = symbol; + lexer_token.type = symbol->ID; + lexer_token.v.symbol = symbol; if(symbol->string != string) { obstack_free(&symbol_obstack, string); } } -#if 0 -static -preprocessor_token_type_t parse_pp_symbol(lexer_t *this) +static void parse_integer_suffix(void) { - do { - obstack_1grow(&symbol_obstack, this->c); - next_char(this); - } while(is_ident_char(this->c)); - obstack_1grow(&symbol_obstack, '\0'); - - char *string = obstack_finish(&symbol_obstack); - symbol_t *symbol = preprocessor_symbol_table_find(string); - obstack_free(&symbol_obstack, string); - - if(symbol == 0) - return TP_ERROR; - - return symbol->ID; + if(c == 'U' || c == 'U') { + /* TODO do something with the suffixes... */ + next_char(); + if(c == 'L' || c == 'l') { + next_char(); + if(c == 'L' || c == 'l') { + next_char(); + } + } + } else if(c == 'l' || c == 'L') { + next_char(); + if(c == 'l' || c == 'L') { + next_char(); + if(c == 'u' || c == 'U') { + next_char(); + } + } else if(c == 'u' || c == 'U') { + next_char(); + } + } } -#endif -static -void parse_number_hex(lexer_t *this, token_t *token) +static void parse_floating_suffix(void) { - assert(this->c == 'x' || this->c == 'X'); - next_char(this); + switch(c) { + /* TODO: do something usefull with the suffixes... */ + case 'f': + case 'F': + case 'l': + case 'L': + next_char(); + break; + default: + break; + } +} - if (!isdigit(this->c) && - !('A' <= this->c && this->c <= 'F') && - !('a' <= this->c && this->c <= 'f')) { - parse_error(this, "premature end of hex number literal"); - token->type = T_ERROR; +static void parse_number_hex(void) +{ + assert(c == 'x' || c == 'X'); + next_char(); + + if (!isdigit(c) && + !('A' <= c && c <= 'F') && + !('a' <= c && c <= 'f')) { + parse_error("premature end of hex number literal"); + lexer_token.type = T_ERROR; return; } int value = 0; - for(;;) { - if (isdigit(this->c)) { - value = 16 * value + this->c - '0'; - } else if ('A' <= this->c && this->c <= 'F') { - value = 16 * value + this->c - 'A' + 10; - } else if ('a' <= this->c && this->c <= 'f') { - value = 16 * value + this->c - 'a' + 10; + while(1) { + if (isdigit(c)) { + value = 16 * value + c - '0'; + } else if ('A' <= c && c <= 'F') { + value = 16 * value + c - 'A' + 10; + } else if ('a' <= c && c <= 'f') { + value = 16 * value + c - 'a' + 10; } else { - token->type = T_INTEGER; - token->v.intvalue = value; + parse_integer_suffix(); + + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = value; return; } - next_char(this); + next_char(); + } + + if(c == '.' || c == 'p' || c == 'P') { + next_char(); + panic("Hex floating point numbers not implemented yet"); } } -static -void parse_number_oct(lexer_t *this, token_t *token) +static void parse_number_oct(void) { - assert(this->c == 'o' || this->c == 'O'); - next_char(this); - int value = 0; - for(;;) { - if ('0' <= this->c && this->c <= '7') { - value = 8 * value + this->c - '0'; - } else { - token->type = T_INTEGER; - token->v.intvalue = value; - return; - } - next_char(this); + while(c >= '0' && c <= '7') { + value = 8 * value + c - '0'; + next_char(); } + if (c == '8' || c == '9') { + parse_error("invalid octal number"); + lexer_token.type = T_ERROR; + return; + } + + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = value; + + parse_integer_suffix(); } -static -void parse_number_dec(lexer_t *this, token_t *token, int first_char) +static void parse_floatingpoint_exponent(long double value) { - int value = 0; - if(first_char > 0) { - assert(first_char >= '0' && first_char <= '9'); - value = first_char - '0'; + unsigned int expo = 0; + long double factor = 10.; + + if(c == '-') { + next_char(); + factor = 0.1; + } else if(c == '+') { + next_char(); } - for(;;) { - if (isdigit(this->c)) { - value = 10 * value + this->c - '0'; - } else { - token->type = T_INTEGER; - token->v.intvalue = value; - return; - } - next_char(this); + while(c >= '0' && c <= '9') { + expo = 10 * expo + (c - '0'); + next_char(); + } + + while(1) { + if(expo & 1) + value *= factor; + expo >>= 1; + if(expo == 0) + break; + factor *= factor; } + + lexer_token.type = T_FLOATINGPOINT; + lexer_token.v.floatvalue = value; + + parse_floating_suffix(); } -static -void parse_number(lexer_t *this, token_t *token) +static void parse_floatingpoint_fract(int integer_part) { - // TODO check for overflow - // TODO check for various invalid inputs sequences + long double value = integer_part; + long double factor = 1.; + + while(c >= '0' && c <= '9') { + factor *= 0.1; + value += (c - '0') * factor; + next_char(); + } - if (this->c == '0') { - next_char(this); - switch (this->c) { + if(c == 'e' || c == 'E') { + next_char(); + parse_floatingpoint_exponent(value); + return; + } + + lexer_token.type = T_FLOATINGPOINT; + lexer_token.v.floatvalue = value; + + parse_floating_suffix(); +} + +static void parse_number_dec(void) +{ + int value = 0; + + while(isdigit(c)) { + value = 10 * value + c - '0'; + next_char(); + } + + if(c == '.') { + next_char(); + parse_floatingpoint_fract(value); + return; + } + if(c == 'e' || c == 'E') { + next_char(); + parse_floatingpoint_exponent(value); + return; + } + parse_integer_suffix(); + + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = value; +} + +static void parse_number(void) +{ + if (c == '0') { + next_char(); + switch (c) { case 'X': - case 'x': parse_number_hex(this, token); break; - case 'o': - case 'O': parse_number_oct(this, token); break; - default: parse_number_dec(this, token, '0'); + case 'x': + parse_number_hex(); + break; + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + parse_number_oct(); + break; + case '.': + next_char(); + parse_floatingpoint_fract(0); + break; + case 'e': + case 'E': + parse_floatingpoint_exponent(0); + break; + case '8': + case '9': + next_char(); + parse_error("invalid octal number"); + lexer_token.type = T_ERROR; + return; + default: + put_back(c); + c = '0'; + parse_number_dec(); + return; } } else { - parse_number_dec(this, token, 0); + parse_number_dec(); } } -static -int parse_escape_sequence(lexer_t *this) +static int parse_octal_sequence(void) { + int value = 0; while(1) { - int c = this->c; - next_char(this); + if(c < '0' || c > '7') + break; + value = 8 * value + c - '0'; + next_char(); + } - switch(c) { - case '"': return '"'; - case '\'': return'\''; - case '\\': - EAT_NEWLINE(break;) - return '\\'; - case 'a': return '\a'; - case 'b': return '\b'; - case 'f': return '\f'; - case 'n': return '\n'; - case 'r': return '\r'; - case 't': return '\t'; - case 'v': return '\v'; - case 'x': /* TODO parse hex number ... */ - parse_error(this, "hex escape sequences not implemented yet"); - return EOF; - case 0 ... 8: /* TODO parse octal number ... */ - parse_error(this, "octal escape sequences not implemented yet"); - return EOF; - case '?': - if(this->c != '?') { - return '?'; - } - /* might be a trigraph */ - next_char(this); - if(replace_trigraph(this)) { - break; - } - put_back(this, this->c); - this->c = '?'; - return '?'; + return value; +} - case EOF: - parse_error(this, "reached end of file while parsing escape sequence"); - return EOF; - default: - parse_error(this, "unknown escape sequence"); - return EOF; +static int parse_hex_sequence(void) +{ + int value = 0; + while(1) { + if (c >= '0' && c <= '9') { + value = 16 * value + c - '0'; + } else if ('A' <= c && c <= 'F') { + value = 16 * value + c - 'A' + 10; + } else if ('a' <= c && c <= 'f') { + value = 16 * value + c - 'a' + 10; + } else { + break; } + next_char(); + } + + return value; +} + +static int parse_escape_sequence(void) +{ + eat('\\'); + + int ec = c; + next_char(); + + switch(ec) { + case '"': return '"'; + case '\'': return'\''; + case '\\': return '\\'; + case '?': return '\?'; + case 'a': return '\a'; + case 'b': return '\b'; + case 'f': return '\f'; + case 'n': return '\n'; + case 'r': return '\r'; + case 't': return '\t'; + case 'v': return '\v'; + case 'x': + return parse_hex_sequence(); + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + return parse_octal_sequence(); + case EOF: + parse_error("reached end of file while parsing escape sequence"); + return EOF; + default: + parse_error("unknown escape sequence"); + return EOF; + } +} + +const char *concat_strings(const char *s1, const char *s2) +{ + size_t len1 = strlen(s1); + size_t len2 = strlen(s2); + + char *concat = obstack_alloc(&symbol_obstack, len1 + len2 + 1); + memcpy(concat, s1, len1); + memcpy(concat + len1, s2, len2 + 1); + + const char *result = strset_insert(&stringset, concat); + if(result != concat) { + obstack_free(&symbol_obstack, concat); } + + return result; } -static -void parse_string_literal(lexer_t *this, token_t *token) +static void parse_string_literal(void) { - unsigned start_linenr = this->source_position.linenr; + unsigned start_linenr = lexer_token.source_position.linenr; char *string; const char *result; - assert(this->c == '"'); - next_char(this); + assert(c == '"'); + next_char(); + int tc; while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - obstack_1grow(&symbol_obstack, '?'); - next_char(this); - break; - ) - + switch(c) { case '\\': - next_char(this); - EAT_NEWLINE(break;) - int c = parse_escape_sequence(this); - obstack_1grow(&symbol_obstack, c); + tc = parse_escape_sequence(); + obstack_1grow(&symbol_obstack, tc); break; case EOF: - error_prefix_at(this, this->source_position.input_name, + error_prefix_at(lexer_token.source_position.input_name, start_linenr); fprintf(stderr, "string has no end\n"); - token->type = T_ERROR; + lexer_token.type = T_ERROR; return; case '"': - next_char(this); + next_char(); goto end_of_string; default: - obstack_1grow(&symbol_obstack, this->c); - next_char(this); + obstack_1grow(&symbol_obstack, c); + next_char(); break; } } @@ -394,267 +608,233 @@ end_of_string: string = obstack_finish(&symbol_obstack); /* check if there is already a copy of the string */ - result = strset_insert(&this->stringset, string); + result = strset_insert(&stringset, string); if(result != string) { obstack_free(&symbol_obstack, string); } - token->type = T_STRING_LITERAL; - token->v.string = result; + lexer_token.type = T_STRING_LITERAL; + lexer_token.v.string = result; } -#define MATCH_NEWLINE(code) \ - case '\r': \ - next_char(this); \ - if(this->c == '\n') { \ - next_char(this); \ - } \ - this->source_position.linenr++; \ - code; \ - case '\n': \ - next_char(this); \ - this->source_position.linenr++; \ - code; - -static -void parse_character_constant(lexer_t *this, token_t *token) +static void parse_character_constant(void) { - assert(this->c == '\''); - next_char(this); + eat('\''); int found_char = 0; while(1) { - switch(this->c) { - SKIP_TRIGRAPHS( - found_char = '?'; - break; - ) - + switch(c) { case '\\': - next_char(this); - EAT_NEWLINE(break;) - found_char = '\\'; + found_char = parse_escape_sequence(); break; MATCH_NEWLINE( - parse_error(this, "newline while parsing character constant"); + parse_error("newline while parsing character constant"); break; ) case '\'': - next_char(this); + next_char(); goto end_of_char_constant; case EOF: - parse_error(this, "EOF while parsing character constant"); - token->type = T_ERROR; + parse_error("EOF while parsing character constant"); + lexer_token.type = T_ERROR; return; default: if(found_char != 0) { - parse_error(this, "more than 1 characters in character " + parse_error("more than 1 characters in character " "constant"); goto end_of_char_constant; } else { - found_char = this->c; - next_char(this); + found_char = c; + next_char(); } break; } } end_of_char_constant: - token->type = T_INTEGER; - token->v.intvalue = found_char; + lexer_token.type = T_INTEGER; + lexer_token.v.intvalue = found_char; } -static -void skip_multiline_comment(lexer_t *this) +static void skip_multiline_comment(void) { - unsigned start_linenr = this->source_position.linenr; - int had_star = 0; + unsigned start_linenr = lexer_token.source_position.linenr; while(1) { - switch(this->c) { + switch(c) { case '*': - next_char(this); - had_star = 1; - break; - - case '/': - next_char(this); - if(had_star) { + next_char(); + if(c == '/') { + next_char(); return; } - had_star = 0; - break; - - case '\\': - next_char(this); - EAT_NEWLINE(break;) - had_star = 0; break; - case '?': - next_char(this); - if(this->c != '?') { - had_star = 0; - break; - } - next_char(this); - if(replace_trigraph(this)) - break; - put_back(this, this->c); - this->c = '?'; - had_star = 0; - /* we don't put back the 2nd ? as the comment text is discarded - * anyway */ - break; - - MATCH_NEWLINE(had_star = 0; break;) + MATCH_NEWLINE(break;) case EOF: - error_prefix_at(this, this->source_position.input_name, + error_prefix_at(lexer_token.source_position.input_name, start_linenr); fprintf(stderr, "at end of file while looking for comment end\n"); return; + default: - had_star = 0; - next_char(this); + next_char(); break; } } } -static -void skip_line_comment(lexer_t *this) +static void skip_line_comment(void) { while(1) { - switch(this->c) { - case '?': - next_char(this); - if(this->c != '?') - break; - next_char(this); - if(replace_trigraph(this)) - break; - put_back(this, '?'); - /* we don't put back the 2nd ? as the comment text is discarded - * anyway */ - break; - - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - } - break; - + switch(c) { case EOF: - case '\r': + return; + case '\n': + case '\r': return; default: - next_char(this); + next_char(); break; } } } -static -void parse_preprocessor_directive(lexer_t *this, token_t *result_token) +static token_t pp_token; + +static inline void next_pp_token(void) { - (void) result_token; - /* skip whitespaces */ - while(this->c == ' ' || this->c == '\t' || this->c == '\r') { - next_char(this); - } + lexer_next_preprocessing_token(); + pp_token = lexer_token; } -void preprocessor_next_token(lexer_t *this, token_t *token) +static void eat_until_newline(void) { - /* skip whitespaces */ - while(this->c == ' ' || this->c == '\t' || this->c == '\r') { - next_char(this); + while(pp_token.type != '\n' && pp_token.type != T_EOF) { + next_pp_token(); } +} - switch(this->c) { - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, token); +static void error_directive(void) +{ + error_prefix(); + fprintf(stderr, "#error directive: \n"); + + /* parse pp-tokens until new-line */ +} + +static void define_directive(void) +{ + lexer_next_preprocessing_token(); + if(lexer_token.type != T_IDENTIFIER) { + parse_error("expected identifier after #define\n"); + eat_until_newline(); } } -void lexer_next_token(lexer_t *this, token_t *token) +static void ifdef_directive(int is_ifndef) { - int line_begin = 0; + (void) is_ifndef; + lexer_next_preprocessing_token(); + //expect_identifier(); + //extect_newline(); +} - while(1) { - switch(this->c) { - case ' ': - case '\t': - next_char(this); - break; +static void endif_directive(void) +{ + //expect_newline(); +} - MATCH_NEWLINE(break;) +static void parse_line_directive(void) +{ + if(pp_token.type != T_INTEGER) { + parse_error("expected integer"); + } else { + lexer_token.source_position.linenr = pp_token.v.intvalue - 1; + next_pp_token(); + } + if(pp_token.type == T_STRING_LITERAL) { + lexer_token.source_position.input_name = pp_token.v.string; + next_pp_token(); + } - case 'A' ... 'Z': - case 'a' ... 'z': - case '_': - parse_symbol(this, token); - return; + eat_until_newline(); +} - case '0' ... '9': - parse_number(this, token); - return; +static void parse_preprocessor_identifier(void) +{ + assert(pp_token.type == T_IDENTIFIER); + symbol_t *symbol = pp_token.v.symbol; - case '"': - parse_string_literal(this, token); - return; + switch(symbol->pp_ID) { + case TP_include: + printf("include - enable header name parsing!\n"); + break; + case TP_define: + define_directive(); + break; + case TP_ifdef: + ifdef_directive(0); + break; + case TP_ifndef: + ifdef_directive(1); + break; + case TP_endif: + endif_directive(); + break; + case TP_line: + next_pp_token(); + parse_line_directive(); + break; + case TP_if: + case TP_else: + case TP_elif: + case TP_undef: + case TP_error: + error_directive(); + break; + case TP_pragma: + break; + } +} - case '\'': - parse_character_constant(this, token); - return; +static void parse_preprocessor_directive(void) +{ + next_pp_token(); - case '\\': - next_char(this); - if(this->c == '\n') { - next_char(this); - this->source_position.linenr++; - break; - } else { - parse_error(this, "unexpected '\\' found"); - token->type = T_ERROR; - } - return; + switch(pp_token.type) { + case T_IDENTIFIER: + parse_preprocessor_identifier(); + break; + case T_INTEGER: + parse_line_directive(); + break; + default: + parse_error("invalid preprocessor directive"); + eat_until_newline(); + break; + } +} #define MAYBE_PROLOG \ - next_char(this); \ + next_char(); \ while(1) { \ - switch(this->c) { + switch(c) { #define MAYBE(ch, set_type) \ case ch: \ - next_char(this); \ - token->type = set_type; \ + next_char(); \ + lexer_token.type = set_type; \ return; #define ELSE_CODE(code) \ - SKIP_TRIGRAPHS( \ - code; \ - ) \ - \ - case '\\': \ - next_char(this); \ - if(this->c == '\n') { \ - next_char(this); \ - this->source_position.linenr++; \ - break; \ - } \ - /* fallthrough */ \ default: \ code; \ } \ @@ -663,19 +843,55 @@ void lexer_next_token(lexer_t *this, token_t *token) #define ELSE(set_type) \ ELSE_CODE( \ - token->type = set_type; \ + lexer_token.type = set_type; \ return; \ ) +void lexer_next_preprocessing_token(void) +{ + while(1) { + switch(c) { + case ' ': + case '\t': + next_char(); + break; + + MATCH_NEWLINE( + lexer_token.type = '\n'; + return; + ) + + SYMBOL_CHARS + parse_symbol(); + /* might be a wide string ( L"string" ) */ + if(c == '"' && (lexer_token.type == T_IDENTIFIER && + lexer_token.v.symbol == symbol_L)) { + parse_string_literal(); + return; + } + return; + + DIGITS + parse_number(); + return; + + case '"': + parse_string_literal(); + return; + + case '\'': + parse_character_constant(); + return; + case '.': MAYBE_PROLOG case '.': MAYBE_PROLOG MAYBE('.', T_DOTDOTDOT) ELSE_CODE( - put_back(this, this->c); - this->c = '.'; - token->type = '.'; + put_back(c); + c = '.'; + lexer_token.type = '.'; return; ) ELSE('.') @@ -695,6 +911,7 @@ void lexer_next_token(lexer_t *this, token_t *token) ELSE('+') case '-': MAYBE_PROLOG + MAYBE('>', T_MINUSGREATER) MAYBE('-', T_MINUSMINUS) MAYBE('=', T_MINUSEQUAL) ELSE('-') @@ -706,14 +923,14 @@ void lexer_next_token(lexer_t *this, token_t *token) MAYBE_PROLOG MAYBE('=', T_SLASHEQUAL) case '*': - next_char(this); - skip_multiline_comment(this); - lexer_next_token(this, token); + next_char(); + skip_multiline_comment(); + lexer_next_preprocessing_token(); return; case '/': - next_char(this); - skip_line_comment(this); - lexer_next_token(this, token); + next_char(); + skip_line_comment(); + lexer_next_preprocessing_token(); return; ELSE('/') case '%': @@ -726,9 +943,9 @@ void lexer_next_token(lexer_t *this, token_t *token) MAYBE_PROLOG MAYBE(':', T_PERCENTCOLONPERCENTCOLON) ELSE_CODE( - put_back(this, this->c); - this->c = '%'; - token->type = T_PERCENTCOLON; + put_back(c); + c = '%'; + lexer_token.type = T_PERCENTCOLON; return; ) ELSE(T_PERCENTCOLON) @@ -737,6 +954,7 @@ void lexer_next_token(lexer_t *this, token_t *token) MAYBE_PROLOG MAYBE(':', T_LESSCOLON) MAYBE('%', T_LESSPERCENT) + MAYBE('=', T_LESSEQUAL) case '<': MAYBE_PROLOG MAYBE('=', T_LESSLESSEQUAL) @@ -744,6 +962,7 @@ void lexer_next_token(lexer_t *this, token_t *token) ELSE('<') case '>': MAYBE_PROLOG + MAYBE('=', T_GREATEREQUAL) case '>': MAYBE_PROLOG MAYBE('=', T_GREATERGREATEREQUAL) @@ -769,33 +988,9 @@ void lexer_next_token(lexer_t *this, token_t *token) case '#': MAYBE_PROLOG MAYBE('#', T_HASHHASH) - ELSE_CODE( - if(line_begin) { - parse_preprocessor_directive(this, token); - return; - } else { - token->type = '#'; - return; - } - ) + ELSE('#') case '?': - next_char(this); - /* just a simple ? */ - if(this->c != '?') { - token->type = '?'; - return; - } - /* might be a trigraph */ - next_char(this); - if(replace_trigraph(this)) { - break; - } - put_back(this, this->c); - this->c = '?'; - token->type = '?'; - return; - case '[': case ']': case '(': @@ -805,47 +1000,69 @@ void lexer_next_token(lexer_t *this, token_t *token) case '~': case ';': case ',': - token->type = this->c; - next_char(this); + case '\\': + lexer_token.type = c; + next_char(); return; case EOF: - token->type = T_EOF; + lexer_token.type = T_EOF; return; default: - next_char(this); - error_prefix(this); - fprintf(stderr, "unknown character '%c' found\n", this->c); - token->type = T_ERROR; + next_char(); + error_prefix(); + fprintf(stderr, "unknown character '%c' found\n", c); + lexer_token.type = T_ERROR; return; } } } -void lexer_init(lexer_t *this, FILE *stream, const char *input_name) +void lexer_next_token(void) { - memset(this, 0, sizeof(this[0])); + lexer_next_preprocessing_token(); + if(lexer_token.type != '\n') + return; - this->input = stream; +newline_found: + do { + lexer_next_preprocessing_token(); + } while(lexer_token.type == '\n'); + + if(lexer_token.type == '#') { + parse_preprocessor_directive(); + goto newline_found; + } +} + +void init_lexer(void) +{ + strset_init(&stringset); +} + +void lexer_open_stream(FILE *stream, const char *input_name) +{ + input = stream; + lexer_token.source_position.linenr = 0; + lexer_token.source_position.input_name = input_name; - this->source_position.linenr = 0; - this->source_position.input_name = input_name; - strset_init(&this->stringset); + symbol_L = symbol_table_insert("L"); - /* we place a virtual '\n' at the beginning so the lexer knows we're at the - * beginning of a line */ - this->c = '\n'; + /* place a virtual \n at the beginning so the lexer knows that we're + * at the beginning of a line */ + c = '\n'; } -void lexer_destroy(lexer_t *this) +void exit_lexer(void) { - (void) this; + strset_destroy(&stringset); } static __attribute__((unused)) void dbg_pos(const source_position_t source_position) { - fprintf(stdout, "%s:%d\n", source_position.input_name, source_position.linenr); + fprintf(stdout, "%s:%d\n", source_position.input_name, + source_position.linenr); fflush(stdout); }