diff options
Diffstat (limited to 'gcc/c-lex.c')
| -rwxr-xr-x | gcc/c-lex.c | 324 |
1 files changed, 25 insertions, 299 deletions
diff --git a/gcc/c-lex.c b/gcc/c-lex.c index c82fc96..bfb2552 100755 --- a/gcc/c-lex.c +++ b/gcc/c-lex.c @@ -30,17 +30,10 @@ Boston, MA 02111-1307, USA. */ #include "c-tree.h" #include "flags.h" #include "c-parse.h" -#include "c-pragma.h" #include "toplev.h" -#ifdef MULTIBYTE_CHARS -#include "mbchar.h" -#include <locale.h> -#endif /* MULTIBYTE_CHARS */ - -#include "cpplib.h" -extern cpp_reader parse_in; -extern cpp_options parse_options; +/* Stream for reading from the input file. */ +FILE *finput; extern void yyprint (FILE *, int, YYSTYPE); @@ -52,12 +45,8 @@ tree ridpointers[(int) RID_MAX]; /* Cause the `yydebug' variable to be defined. */ #define YYDEBUG 1 -extern unsigned char *yy_cur, *yy_lim; - -extern int yy_get_token (); - -#define GETC() (yy_cur < yy_lim ? *yy_cur++ : yy_get_token ()) -#define UNGETC(c) ((void)(c), yy_cur--) +#define GETC() getc(finput) +#define UNGETC(c) ungetc(c, finput) /* the declaration found for the last IDENTIFIER token read in. yylex must look this up to detect typedefs, which get token type TYPENAME, @@ -90,10 +79,6 @@ static int indent_level = 0; /* Number of { minus number of }. */ /* Nonzero if end-of-file has been seen on input. */ static int end_of_file; - -#ifdef HANDLE_GENERIC_PRAGMAS -static int handle_generic_pragma (int); -#endif /* HANDLE_GENERIC_PRAGMAS */ static int whitespace_cr (int); static int skip_white_space (int); static int skip_white_space_on_line (void); @@ -155,17 +140,17 @@ char * init_parse (filename) char *filename; { - parse_in.show_column = 1; - if (! cpp_start_read (&parse_in, filename)) - abort (); - + /* Open input file. */ if (filename == 0 || !strcmp (filename, "-")) - filename = "stdin"; + { + finput = stdin; + filename = "stdin"; + } + else + finput = fopen (filename, "r"); - /* cpp_start_read always puts at least one line directive into the - token buffer. We must arrange to read it out here. */ - yy_cur = parse_in.token_buffer; - yy_lim = CPP_PWRITTEN (&parse_in); + if (finput == 0) + pfatal_with_name (filename); init_lex (); @@ -175,7 +160,7 @@ init_parse (filename) void finish_parse () { - cpp_finish (&parse_in); + fclose(finput); } void @@ -188,12 +173,6 @@ init_lex () and will increment it to 1. */ lineno = 0; -#ifdef MULTIBYTE_CHARS - /* Change to the native locale for multibyte conversions. */ - setlocale (LC_CTYPE, ""); - literal_codeset = getenv ("LANG"); -#endif - maxtoken = 40; token_buffer = (char *) xmalloc (maxtoken + 2); @@ -361,16 +340,12 @@ skip_white_space (c) } } -/* Skips all of the white space at the current location in the input file. - Must use and reset nextchar if it has the next character. */ +/* Skips all of the white space at the current location in the input file. */ void position_after_white_space () { - register int c; - - c = GETC(); - + int c = GETC(); UNGETC (skip_white_space (c)); } @@ -422,27 +397,13 @@ extend_token_buffer (p) return token_buffer + offset; } - -#if defined HANDLE_PRAGMA -/* Local versions of these macros, that can be passed as function pointers. */ -static int -pragma_getc () -{ - return GETC(); -} -static void -pragma_ungetc (arg) - int arg; -{ - UNGETC (arg); -} -#endif +/* At the beginning of the file, check for a #line directive indicating + the real name of the file. */ void check_line_directive() { - check_newline (); - yy_cur--; + ungetc(check_newline(), finput); } /* At the beginning of a line, increment the line number @@ -478,7 +439,7 @@ check_newline () /* If a letter follows, then if the word here is `line', skip it and ignore it; otherwise, ignore the line, with an error - if the word isn't `pragma', `ident', `define', or `undef'. */ + if the word isn't `pragma', `define', or `undef'. */ if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z')) { @@ -489,46 +450,9 @@ check_newline () && GETC() == 'g' && GETC() == 'm' && GETC() == 'a' - && ((c = GETC()) == ' ' || c == '\t' || c == '\n' - || whitespace_cr (c) )) + && ((c = GETC()) == ' ' || c == '\t' || c == '\n' || whitespace_cr (c))) { - while (c == ' ' || c == '\t' || whitespace_cr (c)) - c = GETC (); - if (c == '\n') - return c; - -#if defined HANDLE_PRAGMA || defined HANDLE_GENERIC_PRAGMAS - UNGETC (c); - token = yylex (); - if (token != IDENTIFIER) - goto skipline; -#endif /* HANDLE_PRAGMA || HANDLE_GENERIC_PRAGMAS */ - -#ifdef HANDLE_PRAGMA - /* We invoke HANDLE_PRAGMA before HANDLE_GENERIC_PRAGMAS (if - both are defined), in order to give the back end a chance to - override the interpretation of generic style pragmas. */ - - if (TREE_CODE (yylval.ttype) != IDENTIFIER_NODE) - goto skipline; - - if (HANDLE_PRAGMA (pragma_getc, pragma_ungetc, - IDENTIFIER_POINTER (yylval.ttype))) - return GETC (); -#endif /* HANDLE_PRAGMA */ - -#ifdef HANDLE_GENERIC_PRAGMAS - if (handle_generic_pragma (token)) - return GETC (); -#endif /* HANDLE_GENERIC_PRAGMAS */ - - /* Issue a warning message if we have been asked to do so. - Ignoring unknown pragmas in system header file unless - an explcit -Wunknown-pragmas has been given. */ - if (warn_unknown_pragmas > 1 - || (warn_unknown_pragmas && ! in_system_header)) - warning ("ignoring pragma: %s", token_buffer); - + warning ("ignoring pragma"); goto skipline; } } @@ -568,45 +492,6 @@ check_newline () && ((c = GETC()) == ' ' || c == '\t')) goto linenum; } - else if (c == 'i') - { - if (GETC() == 'd' - && GETC() == 'e' - && GETC() == 'n' - && GETC() == 't' - && ((c = GETC()) == ' ' || c == '\t')) - { - /* #ident. The pedantic warning is now in cccp.c. */ - - /* Here we have just seen `#ident '. - A string constant should follow. */ - - c = skip_white_space_on_line (); - - /* If no argument, ignore the line. */ - if (c == '\n') - return c; - - UNGETC (c); - token = yylex (); - if (token != STRING - || TREE_CODE (yylval.ttype) != STRING_CST) - { - error ("invalid #ident"); - goto skipline; - } - - if (!flag_no_ident) - { -#ifdef ASM_OUTPUT_IDENT - ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (yylval.ttype)); -#endif - } - - /* Skip the rest of this line. */ - goto skipline; - } - } error ("undefined or invalid # directive"); goto skipline; @@ -782,50 +667,7 @@ linenum: c = GETC(); return c; } - -#ifdef HANDLE_GENERIC_PRAGMAS -/* Handle a #pragma directive. - TOKEN is the token we read after `#pragma'. Processes the entire input - line and return non-zero iff the pragma has been successfully parsed. */ - -/* This function has to be in this file, in order to get at - the token types. */ - -static int -handle_generic_pragma (token) - register int token; -{ - register int c; - - for (;;) - { - switch (token) - { - case IDENTIFIER: - case TYPENAME: - case STRING: - case CONSTANT: - handle_pragma_token (token_buffer, yylval.ttype); - break; - default: - handle_pragma_token (token_buffer, NULL); - } - c = GETC (); - - while (c == ' ' || c == '\t') - c = GETC (); - UNGETC (c); - - if (c == '\n' || c == EOF) - return handle_pragma_token (NULL, NULL); - - token = yylex (); - } -} - -#endif /* HANDLE_GENERIC_PRAGMAS */ - #define ENDFILE -1 /* token that represents end-of-file */ /* Read an escape sequence, returning its equivalent as a character, @@ -1022,7 +864,7 @@ yylex () register int value; int wide_flag = 0; - c = GETC(); + c = GETC(); /* Effectively do c = skip_white_space (c) but do it faster in the usual cases. */ @@ -1097,20 +939,10 @@ yylex () case 'u': case 'v': case 'w': case 'x': case 'y': case 'z': case '_': - case '$': letter: p = token_buffer; - while (ISALNUM (c) || c == '_' || c == '$') + while (ISALNUM (c) || c == '_') { - /* Make sure this char really belongs in an identifier. */ - if (c == '$') - { - if (! dollars_in_ident) - error ("`$' in identifier"); - else if (pedantic) - pedwarn ("`$' in identifier"); - } - if (p >= token_buffer + maxtoken) p = extend_token_buffer (p); @@ -1714,7 +1546,7 @@ yylex () UNGETC (c); *p = 0; - if (ISALNUM (c) || c == '.' || c == '_' || c == '$' + if (ISALNUM (c) || c == '.' || c == '_' || (!flag_traditional && (c == '-' || c == '+') && (p[-1] == 'e' || p[-1] == 'E'))) error ("missing white space after number `%s'", token_buffer); @@ -1730,10 +1562,6 @@ yylex () int chars_seen = 0; unsigned width = TYPE_PRECISION (char_type_node); int max_chars; -#ifdef MULTIBYTE_CHARS - int longest_char = local_mb_cur_max (); - (void) local_mbtowc (NULL, NULL, 0); -#endif max_chars = TYPE_PRECISION (integer_type_node) / width; if (wide_flag) @@ -1757,10 +1585,6 @@ yylex () if (width < HOST_BITS_PER_INT && (unsigned) c >= ((unsigned)1 << width)) pedwarn ("escape sequence out of range for character"); -#ifdef MAP_CHARACTER - if (ISPRINT (c)) - c = MAP_CHARACTER (c); -#endif } else if (c == '\n') { @@ -1768,66 +1592,6 @@ yylex () pedwarn ("ANSI C forbids newline in character constant"); lineno++; } - else - { -#ifdef MULTIBYTE_CHARS - wchar_t wc; - int i; - int char_len = -1; - for (i = 1; i <= longest_char; ++i) - { - if (i > maxtoken - 4) - extend_token_buffer (token_buffer); - - token_buffer[i] = c; - char_len = local_mbtowc (& wc, - token_buffer + 1, - i); - if (char_len != -1) - break; - c = GETC (); - } - if (char_len > 1) - { - /* mbtowc sometimes needs an extra char before accepting */ - if (char_len < i) - UNGETC (c); - if (! wide_flag) - { - /* Merge character into result; ignore excess chars. */ - for (i = 1; i <= char_len; ++i) - { - if (i > max_chars) - break; - if (width < HOST_BITS_PER_INT) - result = (result << width) - | (token_buffer[i] - & ((1 << width) - 1)); - else - result = token_buffer[i]; - } - num_chars += char_len; - goto tryagain; - } - c = wc; - } - else - { - if (char_len == -1) - warning ("Ignoring invalid multibyte character"); - if (wide_flag) - c = wc; -#ifdef MAP_CHARACTER - else - c = MAP_CHARACTER (c); -#endif - } -#else /* ! MULTIBYTE_CHARS */ -#ifdef MAP_CHARACTER - c = MAP_CHARACTER (c); -#endif -#endif /* ! MULTIBYTE_CHARS */ - } if (wide_flag) { @@ -1894,10 +1658,6 @@ yylex () { unsigned width = wide_flag ? WCHAR_TYPE_SIZE : TYPE_PRECISION (char_type_node); -#ifdef MULTIBYTE_CHARS - int longest_char = local_mb_cur_max (); - (void) local_mbtowc (NULL, NULL, 0); -#endif c = GETC (); p = token_buffer + 1; @@ -1919,40 +1679,6 @@ yylex () pedwarn ("ANSI C forbids newline in string constant"); lineno++; } - else - { -#ifdef MULTIBYTE_CHARS - wchar_t wc; - int i; - int char_len = -1; - for (i = 0; i < longest_char; ++i) - { - if (p + i >= token_buffer + maxtoken) - p = extend_token_buffer (p); - p[i] = c; - - char_len = local_mbtowc (& wc, p, i + 1); - if (char_len != -1) - break; - c = GETC (); - } - if (char_len == -1) - warning ("Ignoring invalid multibyte character"); - else - { - /* mbtowc sometimes needs an extra char before accepting */ - if (char_len <= i) - UNGETC (c); - if (! wide_flag) - { - p += (i + 1); - c = GETC (); - continue; - } - c = wc; - } -#endif /* MULTIBYTE_CHARS */ - } /* Add this single character into the buffer either as a wchar_t or as a single byte. */ |