X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;f=lexer.c;h=fe51c2ffa176b12abb03470be5ca50ae5a289b35;hb=e7c87251c9a4fd96a0374bd64f0a17068cdf2eaf;hp=b4d50005e96e16753dd78fed3e8f477ab4a1e6a0;hpb=fd16b92f058181088a385c3224c8814831c006d3;p=cparser diff --git a/lexer.c b/lexer.c index b4d5000..fe51c2f 100644 --- a/lexer.c +++ b/lexer.c @@ -21,6 +21,7 @@ #include "diagnostic.h" #include "lexer.h" +#include "symbol_t.h" #include "token_t.h" #include "symbol_table_t.h" #include "adt/error.h" @@ -63,13 +64,28 @@ static strset_t stringset; */ static void parse_error(const char *msg) { - errorf(lexer_token.source_position, "%s", msg); + errorf(&lexer_token.source_position, "%s", msg); +} + +/** + * Prints an internal error message at the current token. + * + * @param msg the error message + */ +static NORETURN internal_error(const char *msg) +{ + internal_errorf(&lexer_token.source_position, "%s", msg); } static inline void next_real_char(void) { assert(bufpos <= bufend); if (bufpos >= bufend) { + if (input == NULL) { + c = EOF; + return; + } + size_t s = fread(buf + MAX_PUTBACK, 1, sizeof(buf) - MAX_PUTBACK, input); if(s == 0) { @@ -332,7 +348,7 @@ static void parse_integer_suffix(bool is_oct_hex) 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) { + } else if(is_oct_hex && v >= 0 && (unsigned long long)v <= (unsigned long long)TARGET_ULONG_MAX) { lexer_token.datatype = type_unsigned_long; return; } @@ -480,7 +496,7 @@ static void parse_number_hex(void) if(c == '.' || c == 'p' || c == 'P') { next_char(); - panic("Hex floating point numbers not implemented yet"); + internal_error("Hex floating point numbers not implemented yet"); } if(*string == '\0') { parse_error("invalid hex number"); @@ -678,7 +694,7 @@ static int digit_value(int digit) { case 'f': case 'F': return 15; default: - panic("wrong character given"); + internal_error("wrong character given"); } } @@ -861,7 +877,7 @@ static void parse_string_literal(void) source_position_t source_position; source_position.input_name = lexer_token.source_position.input_name; source_position.linenr = start_linenr; - errorf(source_position, "string has no end"); + errorf(&source_position, "string has no end"); lexer_token.type = T_ERROR; return; } @@ -906,14 +922,17 @@ end_of_string: */ static void parse_wide_character_constant(void) { + const unsigned start_linenr = lexer_token.source_position.linenr; + eat('\''); - int found_char = 0; while(1) { switch(c) { - case '\\': - found_char = parse_escape_sequence(); + case '\\': { + wchar_rep_t tc = parse_escape_sequence(); + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); break; + } MATCH_NEWLINE( parse_error("newline while parsing character constant"); @@ -924,28 +943,34 @@ static void parse_wide_character_constant(void) next_char(); goto end_of_wide_char_constant; - case EOF: - parse_error("EOF while parsing character constant"); + case EOF: { + source_position_t source_position = lexer_token.source_position; + source_position.linenr = start_linenr; + errorf(&source_position, "EOF while parsing character constant"); lexer_token.type = T_ERROR; return; + } - default: - if(found_char != 0) { - parse_error("more than 1 characters in character " - "constant"); - goto end_of_wide_char_constant; - } else { - found_char = c; - next_char(); - } + default: { + wchar_rep_t tc = (wchar_rep_t) c; + obstack_grow(&symbol_obstack, &tc, sizeof(tc)); + next_char(); break; } + } } -end_of_wide_char_constant: - lexer_token.type = T_INTEGER; - lexer_token.v.intvalue = found_char; - lexer_token.datatype = type_wchar_t; +end_of_wide_char_constant:; + size_t size = (size_t) obstack_object_size(&symbol_obstack); + assert(size % sizeof(wchar_rep_t) == 0); + size /= sizeof(wchar_rep_t); + + const wchar_rep_t *string = obstack_finish(&symbol_obstack); + + lexer_token.type = T_WIDE_CHARACTER_CONSTANT; + lexer_token.v.wide_string.begin = string; + lexer_token.v.wide_string.size = size; + lexer_token.datatype = type_wchar_t; } /** @@ -970,7 +995,7 @@ static void parse_wide_string_literal(void) source_position_t source_position; source_position.input_name = lexer_token.source_position.input_name; source_position.linenr = start_linenr; - errorf(source_position, "string has no end"); + errorf(&source_position, "string has no end"); lexer_token.type = T_ERROR; return; } @@ -1022,32 +1047,32 @@ static void parse_character_constant(void) eat('\''); - int tc; while(1) { switch(c) { - case '\\': - tc = parse_escape_sequence(); + case '\\': { + int tc = parse_escape_sequence(); obstack_1grow(&symbol_obstack, (char) tc); break; + } MATCH_NEWLINE( parse_error("newline while parsing character constant"); break; ) + case '\'': + next_char(); + goto end_of_char_constant; + case EOF: { source_position_t source_position; source_position.input_name = lexer_token.source_position.input_name; source_position.linenr = start_linenr; - errorf(source_position, "EOF while parsing character constant"); + errorf(&source_position, "EOF while parsing character constant"); lexer_token.type = T_ERROR; return; } - case '\'': - next_char(); - goto end_of_char_constant; - default: obstack_1grow(&symbol_obstack, (char) c); next_char(); @@ -1060,7 +1085,7 @@ end_of_char_constant:; const size_t size = (size_t)obstack_object_size(&symbol_obstack); const char *const string = obstack_finish(&symbol_obstack); - lexer_token.type = T_CHARS; + lexer_token.type = T_CHARACTER_CONSTANT; lexer_token.v.string.begin = string; lexer_token.v.string.size = size; lexer_token.datatype = type_int; @@ -1095,7 +1120,7 @@ static void skip_multiline_comment(void) source_position_t source_position; source_position.input_name = lexer_token.source_position.input_name; source_position.linenr = start_linenr; - errorf(source_position, "at end of file while looking for comment end"); + errorf(&source_position, "at end of file while looking for comment end"); return; } @@ -1264,7 +1289,7 @@ static void parse_pragma(void) { if (value != STDC_VALUE_UNKNOWN) { unknown_pragma = false; } else { - errorf(pp_token.source_position, "bad STDC pragma argument"); + errorf(&pp_token.source_position, "bad STDC pragma argument"); } } } @@ -1273,7 +1298,7 @@ static void parse_pragma(void) { } eat_until_newline(); if (unknown_pragma && warning.unknown_pragmas) { - warningf(pp_token.source_position, "encountered unknown #pragma"); + warningf(&pp_token.source_position, "encountered unknown #pragma"); } } @@ -1356,7 +1381,7 @@ static void parse_preprocessor_directive(void) #define ELSE_CODE(code) \ default: \ - code; \ + code \ } \ } /* end of while(1) */ \ break; @@ -1473,25 +1498,25 @@ void lexer_next_preprocessing_token(void) ELSE('/') case '%': MAYBE_PROLOG - MAYBE('>', T_PERCENTGREATER) + MAYBE('>', '}') MAYBE('=', T_PERCENTEQUAL) case ':': MAYBE_PROLOG case '%': MAYBE_PROLOG - MAYBE(':', T_PERCENTCOLONPERCENTCOLON) + MAYBE(':', T_HASHHASH) ELSE_CODE( put_back(c); c = '%'; - lexer_token.type = T_PERCENTCOLON; + lexer_token.type = '#'; return; ) - ELSE(T_PERCENTCOLON) + ELSE('#') ELSE('%') case '<': MAYBE_PROLOG - MAYBE(':', T_LESSCOLON) - MAYBE('%', T_LESSPERCENT) + MAYBE(':', '[') + MAYBE('%', '{') MAYBE('=', T_LESSEQUAL) case '<': MAYBE_PROLOG @@ -1517,7 +1542,7 @@ void lexer_next_preprocessing_token(void) ELSE('|') case ':': MAYBE_PROLOG - MAYBE('>', T_COLONGREATER) + MAYBE('>', ']') ELSE(':') case '=': MAYBE_PROLOG @@ -1549,7 +1574,7 @@ void lexer_next_preprocessing_token(void) default: next_char(); - errorf(lexer_token.source_position, "unknown character '%c' found\n", c); + errorf(&lexer_token.source_position, "unknown character '%c' found\n", c); lexer_token.type = T_ERROR; return; } @@ -1576,6 +1601,7 @@ newline_found: void init_lexer(void) { strset_init(&stringset); + symbol_L = symbol_table_insert("L"); } void lexer_open_stream(FILE *stream, const char *input_name) @@ -1584,7 +1610,6 @@ void lexer_open_stream(FILE *stream, const char *input_name) lexer_token.source_position.linenr = 0; lexer_token.source_position.input_name = input_name; - symbol_L = symbol_table_insert("L"); bufpos = NULL; bufend = NULL; @@ -1593,6 +1618,20 @@ void lexer_open_stream(FILE *stream, const char *input_name) c = '\n'; } +void lexer_open_buffer(const char *buffer, size_t len, const char *input_name) +{ + input = NULL; + lexer_token.source_position.linenr = 0; + lexer_token.source_position.input_name = input_name; + + bufpos = buffer; + bufend = buffer + len; + + /* place a virtual \n at the beginning so the lexer knows that we're + * at the beginning of a line */ + c = '\n'; +} + void exit_lexer(void) { strset_destroy(&stringset);