X-Git-Url: http://git.sourceforge.jp/view?a=blobdiff_plain;f=gcc%2Fc-lex.c;h=47515e7b7032020e8963ce4962b339bf126cf708;hb=ca474d266e650d9bfb9a71574c11b8f49c7d3d62;hp=cf65ceacb2c29aca73a853fdfb389eec8ab645c9;hpb=0f2b564d9b2753303a2eef00fa9ed5f58d85c782;p=pf3gnuchains%2Fgcc-fork.git diff --git a/gcc/c-lex.c b/gcc/c-lex.c index cf65ceacb2c..47515e7b703 100644 --- a/gcc/c-lex.c +++ b/gcc/c-lex.c @@ -1,34 +1,37 @@ -/* Lexical analyzer for C and Objective C. +/* Mainly the interface between cpplib and the C front ends. Copyright (C) 1987, 1988, 1989, 1992, 1994, 1995, 1996, 1997 - 1998, 1999, 2000 Free Software Foundation, Inc. + 1998, 1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc. -This file is part of GNU CC. +This file is part of GCC. -GNU CC is free software; you can redistribute it and/or modify -it under the terms of the GNU General Public License as published by -the Free Software Foundation; either version 2, or (at your option) -any later version. +GCC is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. -GNU CC is distributed in the hope that it will be useful, -but WITHOUT ANY WARRANTY; without even the implied warranty of -MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -GNU General Public License for more details. +GCC is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. You should have received a copy of the GNU General Public License -along with GNU CC; see the file COPYING. If not, write to -the Free Software Foundation, 59 Temple Place - Suite 330, -Boston, MA 02111-1307, USA. */ +along with GCC; see the file COPYING. If not, write to the Free +Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. */ #include "config.h" #include "system.h" +#include "coretypes.h" +#include "tm.h" +#include "real.h" #include "rtl.h" -#include "expr.h" #include "tree.h" +#include "expr.h" #include "input.h" #include "output.h" -#include "c-lex.h" #include "c-tree.h" +#include "c-common.h" #include "flags.h" #include "timevar.h" #include "cpplib.h" @@ -37,74 +40,51 @@ Boston, MA 02111-1307, USA. */ #include "intl.h" #include "tm_p.h" #include "splay-tree.h" +#include "debug.h" -/* MULTIBYTE_CHARS support only works for native compilers. - ??? Ideally what we want is to model widechar support after - the current floating point support. */ -#ifdef CROSS_COMPILE -#undef MULTIBYTE_CHARS -#endif - -#ifdef MULTIBYTE_CHARS -#include "mbchar.h" -#include -#endif /* MULTIBYTE_CHARS */ -#ifndef GET_ENVIRONMENT -#define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME)) -#endif - -/* The original file name, before changing "-" to "stdin". */ -static const char *orig_filename; +/* The current line map. */ +static const struct line_map *map; -/* Private idea of the line number. See discussion in c_lex(). */ -static int lex_lineno; +/* The line used to refresh the lineno global variable after each token. */ +static unsigned int src_lineno; /* We may keep statistics about how long which files took to compile. */ static int header_time, body_time; static splay_tree file_info_tree; -/* Cause the `yydebug' variable to be defined. */ -#define YYDEBUG 1 - -/* File used for outputting assembler code. */ -extern FILE *asm_out_file; - #undef WCHAR_TYPE_SIZE #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node) /* Number of bytes in a wide character. */ #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT) -int indent_level; /* Number of { minus number of }. */ int pending_lang_change; /* If we need to switch languages - C++ only */ int c_header_level; /* depth in C headers - C++ only */ -/* Nonzero tells yylex to ignore \ in string constants. */ -static int ignore_escape_flag; - -static const char *readescape PARAMS ((const char *, const char *, - unsigned int *)); -static const char *read_ucs PARAMS ((const char *, const char *, - unsigned int *, int)); -static void parse_float PARAMS ((PTR)); -static tree lex_number PARAMS ((const char *, unsigned int)); -static tree lex_string PARAMS ((const char *, unsigned int, int)); -static tree lex_charconst PARAMS ((const char *, unsigned int, int)); -static void update_header_times PARAMS ((const char *)); -static int dump_one_header PARAMS ((splay_tree_node, void *)); -static void cb_ident PARAMS ((cpp_reader *, const cpp_string *)); -static void cb_change_file PARAMS ((cpp_reader *, const cpp_file_change *)); -static void cb_def_pragma PARAMS ((cpp_reader *)); +static tree interpret_integer (const cpp_token *, unsigned int); +static tree interpret_float (const cpp_token *, unsigned int); +static enum integer_type_kind + narrowest_unsigned_type (tree, unsigned int); +static enum integer_type_kind + narrowest_signed_type (tree, unsigned int); +static enum cpp_ttype lex_string (const cpp_token *, tree *, bool); +static tree lex_charconst (const cpp_token *); +static void update_header_times (const char *); +static int dump_one_header (splay_tree_node, void *); +static void cb_line_change (cpp_reader *, const cpp_token *, int); +static void cb_dir_change (cpp_reader *, const char *); +static void cb_ident (cpp_reader *, unsigned int, const cpp_string *); +static void cb_def_pragma (cpp_reader *, unsigned int); +static void cb_define (cpp_reader *, unsigned int, cpp_hashnode *); +static void cb_undef (cpp_reader *, unsigned int, cpp_hashnode *); -const char * -init_c_lex (filename) - const char *filename; +void +init_c_lex (void) { + struct cpp_callbacks *cb; struct c_fileinfo *toplevel; - orig_filename = filename; - - /* Set up filename timing. Must happen before cpp_start_read. */ + /* Set up filename timing. Must happen before cpp_read_main_file. */ file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp, 0, (splay_tree_delete_value_fn)free); @@ -115,42 +95,28 @@ init_c_lex (filename) body_time = get_run_time (); toplevel->time = body_time; } - -#ifdef MULTIBYTE_CHARS - /* Change to the native locale for multibyte conversions. */ - setlocale (LC_CTYPE, ""); - GET_ENVIRONMENT (literal_codeset, "LANG"); -#endif - parse_in->cb.ident = cb_ident; - parse_in->cb.change_file = cb_change_file; - parse_in->cb.def_pragma = cb_def_pragma; + cb = cpp_get_callbacks (parse_in); - if (filename == 0 || !strcmp (filename, "-")) - filename = "stdin"; + cb->line_change = cb_line_change; + cb->dir_change = cb_dir_change; + cb->ident = cb_ident; + cb->def_pragma = cb_def_pragma; + cb->valid_pch = c_common_valid_pch; + cb->read_pch = c_common_read_pch; - /* Start it at 0, because check_newline is called at the very beginning - and will increment it to 1. */ - lineno = lex_lineno = 0; - - return filename; -} - -/* A thin wrapper around the real parser that initializes the - integrated preprocessor after debug output has been initialized. */ - -int -yyparse() -{ - if (! cpp_start_read (parse_in, orig_filename)) - return 1; /* cpplib has emitted an error. */ - - return yyparse_1(); + /* Set the debug callbacks if we can use them. */ + if (debug_info_level == DINFO_LEVEL_VERBOSE + && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG + || write_symbols == VMS_AND_DWARF2_DEBUG)) + { + cb->define = cb_define; + cb->undef = cb_undef; + } } struct c_fileinfo * -get_fileinfo (name) - const char *name; +get_fileinfo (const char *name) { splay_tree_node n; struct c_fileinfo *fi; @@ -159,7 +125,7 @@ get_fileinfo (name) if (n) return (struct c_fileinfo *) n->value; - fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo)); + fi = xmalloc (sizeof (struct c_fileinfo)); fi->time = 0; fi->interface_only = 0; fi->interface_unknown = 1; @@ -169,8 +135,7 @@ get_fileinfo (name) } static void -update_header_times (name) - const char *name; +update_header_times (const char *name) { /* Changing files again. This means currently collected time is charged against header time, and body time starts back at 0. */ @@ -185,9 +150,7 @@ update_header_times (name) } static int -dump_one_header (n, dummy) - splay_tree_node n; - void *dummy ATTRIBUTE_UNUSED; +dump_one_header (splay_tree_node n, void *dummy ATTRIBUTE_UNUSED) { print_time ((const char *) n->key, ((struct c_fileinfo *) n->value)->time); @@ -195,7 +158,7 @@ dump_one_header (n, dummy) } void -dump_time_statistics () +dump_time_statistics (void) { struct c_fileinfo *file = get_fileinfo (input_filename); int this_time = get_run_time (); @@ -211,47 +174,63 @@ dump_time_statistics () splay_tree_foreach (file_info_tree, dump_one_header, 0); } -/* Not yet handled: #pragma, #define, #undef. - No need to deal with linemarkers under normal conditions. */ - static void -cb_ident (pfile, str) - cpp_reader *pfile ATTRIBUTE_UNUSED; - const cpp_string *str ATTRIBUTE_UNUSED; +cb_ident (cpp_reader *pfile ATTRIBUTE_UNUSED, + unsigned int line ATTRIBUTE_UNUSED, + const cpp_string *str ATTRIBUTE_UNUSED) { #ifdef ASM_OUTPUT_IDENT if (! flag_no_ident) { /* Convert escapes in the string. */ - tree value = lex_string ((const char *)str->text, str->len, 0); - ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value)); + cpp_string cstr = { 0, 0 }; + if (cpp_interpret_string (pfile, str, 1, &cstr, false)) + { + ASM_OUTPUT_IDENT (asm_out_file, (const char *) cstr.text); + free ((void *)cstr.text); + } } #endif } +/* Called at the start of every non-empty line. TOKEN is the first + lexed token on the line. Used for diagnostic line numbers. */ +static void +cb_line_change (cpp_reader *pfile ATTRIBUTE_UNUSED, const cpp_token *token, + int parsing_args ATTRIBUTE_UNUSED) +{ + src_lineno = SOURCE_LINE (map, token->line); +} + static void -cb_change_file (pfile, fc) - cpp_reader *pfile ATTRIBUTE_UNUSED; - const cpp_file_change *fc; +cb_dir_change (cpp_reader *pfile ATTRIBUTE_UNUSED, const char *dir) { - if (fc->from.filename == 0) - main_input_filename = fc->to.filename; - in_system_header = fc->sysp; + if (! set_src_pwd (dir)) + warning ("too late for # directive to set debug directory"); +} - /* Do the actions implied by the preceding numbers. */ - if (fc->reason == FC_ENTER) +void +fe_file_change (const struct line_map *new_map) +{ + unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line); + + if (new_map->reason == LC_ENTER) { - /* FIXME. Don't stack the main buffer on the input stack. */ - if (fc->from.filename) + /* Don't stack the main buffer on the input stack; + we already did in compile_file. */ + if (map == NULL) + main_input_filename = new_map->to_file; + else { - lineno = lex_lineno; - push_srcloc (fc->to.filename, 1); - input_file_stack->indent_level = indent_level; - debug_start_source_file (fc->to.filename); + int included_at = SOURCE_LINE (new_map - 1, new_map->from_line - 1); + + input_line = included_at; + push_srcloc (new_map->to_file, 1); + (*debug_hooks->start_source_file) (included_at, new_map->to_file); #ifndef NO_IMPLICIT_EXTERN_C if (c_header_level) ++c_header_level; - else if (fc->externc) + else if (new_map->sysp == 2) { c_header_level = 1; ++pending_lang_change; @@ -259,740 +238,186 @@ cb_change_file (pfile, fc) #endif } } - else if (fc->reason == FC_LEAVE) + else if (new_map->reason == LC_LEAVE) { - /* Popping out of a file. */ - if (input_file_stack->next) - { #ifndef NO_IMPLICIT_EXTERN_C - if (c_header_level && --c_header_level == 0) - { - if (fc->externc) - warning ("badly nested C headers from preprocessor"); - --pending_lang_change; - } -#endif -#if 0 - if (indent_level != input_file_stack->indent_level) - { - warning_with_file_and_line - (input_filename, lex_lineno, - "This file contains more '%c's than '%c's.", - indent_level > input_file_stack->indent_level ? '{' : '}', - indent_level > input_file_stack->indent_level ? '}' : '{'); - } -#endif - pop_srcloc (); - debug_end_source_file (input_file_stack->line); + if (c_header_level && --c_header_level == 0) + { + if (new_map->sysp == 2) + warning ("badly nested C headers from preprocessor"); + --pending_lang_change; } - else - error ("leaving more files than we entered"); - } - else if (fc->reason == FC_RENAME) - input_filename = fc->to.filename; +#endif + pop_srcloc (); - update_header_times (fc->to.filename); + (*debug_hooks->end_source_file) (to_line); + } - input_filename = fc->to.filename; - lex_lineno = fc->to.lineno; + update_header_times (new_map->to_file); + in_system_header = new_map->sysp != 0; + input_filename = new_map->to_file; + input_line = to_line; + map = new_map; /* Hook for C++. */ extract_interface_info (); } static void -cb_def_pragma (pfile) - cpp_reader *pfile; +cb_def_pragma (cpp_reader *pfile, unsigned int line) { /* Issue a warning message if we have been asked to do so. Ignore unknown pragmas in system headers unless an explicit - -Wunknown-pragmas has been given. */ + -Wunknown-pragmas has been given. */ if (warn_unknown_pragmas > in_system_header) { - const unsigned char *space, *name = 0; - cpp_token s; - - cpp_get_token (pfile, &s); - space = cpp_token_as_text (pfile, &s); - cpp_get_token (pfile, &s); - if (s.type == CPP_NAME) - name = cpp_token_as_text (pfile, &s); + const unsigned char *space, *name; + const cpp_token *s; - if (name) - warning ("ignoring #pragma %s %s", space, name); - else - warning ("ignoring #pragma %s", space); - } -} - -/* Parse a '\uNNNN' or '\UNNNNNNNN' sequence. - - [lex.charset]: The character designated by the universal-character-name - \UNNNNNNNN is that character whose character short name in ISO/IEC 10646 - is NNNNNNNN; the character designated by the universal-character-name - \uNNNN is that character whose character short name in ISO/IEC 10646 is - 0000NNNN. If the hexadecimal value for a universal character name is - less than 0x20 or in the range 0x7F-0x9F (inclusive), or if the - universal character name designates a character in the basic source - character set, then the program is ill-formed. - - We assume that wchar_t is Unicode, so we don't need to do any - mapping. Is this ever wrong? */ - -static const char * -read_ucs (p, limit, cptr, length) - const char *p; - const char *limit; - unsigned int *cptr; - int length; -{ - unsigned int code = 0; - int c; - - for (; length; --length) - { - if (p >= limit) + space = name = (const unsigned char *) ""; + s = cpp_get_token (pfile); + if (s->type != CPP_EOF) { - error ("incomplete universal-character-name"); - break; + space = cpp_token_as_text (pfile, s); + s = cpp_get_token (pfile); + if (s->type == CPP_NAME) + name = cpp_token_as_text (pfile, s); } - c = *p++; - if (! ISXDIGIT (c)) - { - error ("non hex digit '%c' in universal-character-name", c); - p--; - break; - } - - code <<= 4; - if (c >= 'a' && c <= 'f') - code += c - 'a' + 10; - if (c >= 'A' && c <= 'F') - code += c - 'A' + 10; - if (c >= '0' && c <= '9') - code += c - '0'; - } - -#ifdef TARGET_EBCDIC - sorry ("universal-character-name on EBCDIC target"); - *cptr = 0x3f; /* EBCDIC invalid character */ - return p; -#endif - - if (code > 0x9f && !(code & 0x80000000)) - /* True extended character, OK. */; - else if (code >= 0x20 && code < 0x7f) - { - /* ASCII printable character. The C character set consists of all of - these except $, @ and `. We use hex escapes so that this also - works with EBCDIC hosts. */ - if (code != 0x24 && code != 0x40 && code != 0x60) - error ("universal-character-name used for '%c'", code); + input_line = SOURCE_LINE (map, line); + warning ("ignoring #pragma %s %s", space, name); } - else - error ("invalid universal-character-name"); - - *cptr = code; - return p; } -/* Read an escape sequence and write its character equivalent into *CPTR. - P is the input pointer, which is just after the backslash. LIMIT - is how much text we have. - Returns the updated input pointer. */ - -static const char * -readescape (p, limit, cptr) - const char *p; - const char *limit; - unsigned int *cptr; +/* #define callback for DWARF and DWARF2 debug info. */ +static void +cb_define (cpp_reader *pfile, unsigned int line, cpp_hashnode *node) { - unsigned int c, code, count; - unsigned firstdig = 0; - int nonnull; - - if (p == limit) - { - /* cpp has already issued an error for this. */ - *cptr = 0; - return p; - } - - c = *p++; - - switch (c) - { - case 'x': - if (warn_traditional && !in_system_header) - warning ("the meaning of `\\x' varies with -traditional"); - - if (flag_traditional) - { - *cptr = 'x'; - return p; - } - - code = 0; - count = 0; - nonnull = 0; - while (p < limit) - { - c = *p++; - if (! ISXDIGIT (c)) - { - p--; - break; - } - code *= 16; - if (c >= 'a' && c <= 'f') - code += c - 'a' + 10; - if (c >= 'A' && c <= 'F') - code += c - 'A' + 10; - if (c >= '0' && c <= '9') - code += c - '0'; - if (code != 0 || count != 0) - { - if (count == 0) - firstdig = code; - count++; - } - nonnull = 1; - } - if (! nonnull) - { - warning ("\\x used with no following hex digits"); - *cptr = 'x'; - return p; - } - else if (count == 0) - /* Digits are all 0's. Ok. */ - ; - else if ((count - 1) * 4 >= TYPE_PRECISION (integer_type_node) - || (count > 1 - && (((unsigned)1 - << (TYPE_PRECISION (integer_type_node) - - (count - 1) * 4)) - <= firstdig))) - pedwarn ("hex escape out of range"); - *cptr = code; - return p; - - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': - code = 0; - for (count = 0; count < 3; count++) - { - if (c < '0' || c > '7') - { - p--; - break; - } - code = (code * 8) + (c - '0'); - if (p == limit) - break; - c = *p++; - } - - if (count == 3) - p--; - - *cptr = code; - return p; - - case '\\': case '\'': case '"': case '?': - *cptr = c; - return p; - - case 'n': *cptr = TARGET_NEWLINE; return p; - case 't': *cptr = TARGET_TAB; return p; - case 'r': *cptr = TARGET_CR; return p; - case 'f': *cptr = TARGET_FF; return p; - case 'b': *cptr = TARGET_BS; return p; - case 'v': *cptr = TARGET_VT; return p; - case 'a': - if (warn_traditional && !in_system_header) - warning ("the meaning of '\\a' varies with -traditional"); - *cptr = flag_traditional ? c : TARGET_BELL; - return p; - - /* Warnings and support checks handled by read_ucs(). */ - case 'u': case 'U': - if (c_language != clk_cplusplus && !flag_isoc99) - break; - - if (warn_traditional && !in_system_header) - warning ("the meaning of '\\%c' varies with -traditional", c); - - return read_ucs (p, limit, cptr, c == 'u' ? 4 : 8); - - case 'e': case 'E': - if (pedantic) - pedwarn ("non-ISO-standard escape sequence, '\\%c'", c); - *cptr = TARGET_ESC; return p; - - /* '\(', etc, are used at beginning of line to avoid confusing Emacs. - '\%' is used to prevent SCCS from getting confused. */ - case '(': case '{': case '[': case '%': - if (pedantic) - pedwarn ("unknown escape sequence '\\%c'", c); - *cptr = c; - return p; - } - - if (ISGRAPH (c)) - pedwarn ("unknown escape sequence '\\%c'", c); - else - pedwarn ("unknown escape sequence: '\\' followed by char 0x%x", c); - - *cptr = c; - return p; + (*debug_hooks->define) (SOURCE_LINE (map, line), + (const char *) cpp_macro_definition (pfile, node)); } -#if 0 /* not yet */ -/* Returns nonzero if C is a universal-character-name. Give an error if it - is not one which may appear in an identifier, as per [extendid]. - - Note that extended character support in identifiers has not yet been - implemented. It is my personal opinion that this is not a desirable - feature. Portable code cannot count on support for more than the basic - identifier character set. */ - -static inline int -is_extended_char (c) - int c; +/* #undef callback for DWARF and DWARF2 debug info. */ +static void +cb_undef (cpp_reader *pfile ATTRIBUTE_UNUSED, unsigned int line, + cpp_hashnode *node) { -#ifdef TARGET_EBCDIC - return 0; -#else - /* ASCII. */ - if (c < 0x7f) - return 0; - - /* None of the valid chars are outside the Basic Multilingual Plane (the - low 16 bits). */ - if (c > 0xffff) - { - error ("universal-character-name '\\U%08x' not valid in identifier", c); - return 1; - } - - /* Latin */ - if ((c >= 0x00c0 && c <= 0x00d6) - || (c >= 0x00d8 && c <= 0x00f6) - || (c >= 0x00f8 && c <= 0x01f5) - || (c >= 0x01fa && c <= 0x0217) - || (c >= 0x0250 && c <= 0x02a8) - || (c >= 0x1e00 && c <= 0x1e9a) - || (c >= 0x1ea0 && c <= 0x1ef9)) - return 1; - - /* Greek */ - if ((c == 0x0384) - || (c >= 0x0388 && c <= 0x038a) - || (c == 0x038c) - || (c >= 0x038e && c <= 0x03a1) - || (c >= 0x03a3 && c <= 0x03ce) - || (c >= 0x03d0 && c <= 0x03d6) - || (c == 0x03da) - || (c == 0x03dc) - || (c == 0x03de) - || (c == 0x03e0) - || (c >= 0x03e2 && c <= 0x03f3) - || (c >= 0x1f00 && c <= 0x1f15) - || (c >= 0x1f18 && c <= 0x1f1d) - || (c >= 0x1f20 && c <= 0x1f45) - || (c >= 0x1f48 && c <= 0x1f4d) - || (c >= 0x1f50 && c <= 0x1f57) - || (c == 0x1f59) - || (c == 0x1f5b) - || (c == 0x1f5d) - || (c >= 0x1f5f && c <= 0x1f7d) - || (c >= 0x1f80 && c <= 0x1fb4) - || (c >= 0x1fb6 && c <= 0x1fbc) - || (c >= 0x1fc2 && c <= 0x1fc4) - || (c >= 0x1fc6 && c <= 0x1fcc) - || (c >= 0x1fd0 && c <= 0x1fd3) - || (c >= 0x1fd6 && c <= 0x1fdb) - || (c >= 0x1fe0 && c <= 0x1fec) - || (c >= 0x1ff2 && c <= 0x1ff4) - || (c >= 0x1ff6 && c <= 0x1ffc)) - return 1; - - /* Cyrillic */ - if ((c >= 0x0401 && c <= 0x040d) - || (c >= 0x040f && c <= 0x044f) - || (c >= 0x0451 && c <= 0x045c) - || (c >= 0x045e && c <= 0x0481) - || (c >= 0x0490 && c <= 0x04c4) - || (c >= 0x04c7 && c <= 0x04c8) - || (c >= 0x04cb && c <= 0x04cc) - || (c >= 0x04d0 && c <= 0x04eb) - || (c >= 0x04ee && c <= 0x04f5) - || (c >= 0x04f8 && c <= 0x04f9)) - return 1; - - /* Armenian */ - if ((c >= 0x0531 && c <= 0x0556) - || (c >= 0x0561 && c <= 0x0587)) - return 1; - - /* Hebrew */ - if ((c >= 0x05d0 && c <= 0x05ea) - || (c >= 0x05f0 && c <= 0x05f4)) - return 1; - - /* Arabic */ - if ((c >= 0x0621 && c <= 0x063a) - || (c >= 0x0640 && c <= 0x0652) - || (c >= 0x0670 && c <= 0x06b7) - || (c >= 0x06ba && c <= 0x06be) - || (c >= 0x06c0 && c <= 0x06ce) - || (c >= 0x06e5 && c <= 0x06e7)) - return 1; - - /* Devanagari */ - if ((c >= 0x0905 && c <= 0x0939) - || (c >= 0x0958 && c <= 0x0962)) - return 1; - - /* Bengali */ - if ((c >= 0x0985 && c <= 0x098c) - || (c >= 0x098f && c <= 0x0990) - || (c >= 0x0993 && c <= 0x09a8) - || (c >= 0x09aa && c <= 0x09b0) - || (c == 0x09b2) - || (c >= 0x09b6 && c <= 0x09b9) - || (c >= 0x09dc && c <= 0x09dd) - || (c >= 0x09df && c <= 0x09e1) - || (c >= 0x09f0 && c <= 0x09f1)) - return 1; - - /* Gurmukhi */ - if ((c >= 0x0a05 && c <= 0x0a0a) - || (c >= 0x0a0f && c <= 0x0a10) - || (c >= 0x0a13 && c <= 0x0a28) - || (c >= 0x0a2a && c <= 0x0a30) - || (c >= 0x0a32 && c <= 0x0a33) - || (c >= 0x0a35 && c <= 0x0a36) - || (c >= 0x0a38 && c <= 0x0a39) - || (c >= 0x0a59 && c <= 0x0a5c) - || (c == 0x0a5e)) - return 1; - - /* Gujarati */ - if ((c >= 0x0a85 && c <= 0x0a8b) - || (c == 0x0a8d) - || (c >= 0x0a8f && c <= 0x0a91) - || (c >= 0x0a93 && c <= 0x0aa8) - || (c >= 0x0aaa && c <= 0x0ab0) - || (c >= 0x0ab2 && c <= 0x0ab3) - || (c >= 0x0ab5 && c <= 0x0ab9) - || (c == 0x0ae0)) - return 1; - - /* Oriya */ - if ((c >= 0x0b05 && c <= 0x0b0c) - || (c >= 0x0b0f && c <= 0x0b10) - || (c >= 0x0b13 && c <= 0x0b28) - || (c >= 0x0b2a && c <= 0x0b30) - || (c >= 0x0b32 && c <= 0x0b33) - || (c >= 0x0b36 && c <= 0x0b39) - || (c >= 0x0b5c && c <= 0x0b5d) - || (c >= 0x0b5f && c <= 0x0b61)) - return 1; - - /* Tamil */ - if ((c >= 0x0b85 && c <= 0x0b8a) - || (c >= 0x0b8e && c <= 0x0b90) - || (c >= 0x0b92 && c <= 0x0b95) - || (c >= 0x0b99 && c <= 0x0b9a) - || (c == 0x0b9c) - || (c >= 0x0b9e && c <= 0x0b9f) - || (c >= 0x0ba3 && c <= 0x0ba4) - || (c >= 0x0ba8 && c <= 0x0baa) - || (c >= 0x0bae && c <= 0x0bb5) - || (c >= 0x0bb7 && c <= 0x0bb9)) - return 1; - - /* Telugu */ - if ((c >= 0x0c05 && c <= 0x0c0c) - || (c >= 0x0c0e && c <= 0x0c10) - || (c >= 0x0c12 && c <= 0x0c28) - || (c >= 0x0c2a && c <= 0x0c33) - || (c >= 0x0c35 && c <= 0x0c39) - || (c >= 0x0c60 && c <= 0x0c61)) - return 1; - - /* Kannada */ - if ((c >= 0x0c85 && c <= 0x0c8c) - || (c >= 0x0c8e && c <= 0x0c90) - || (c >= 0x0c92 && c <= 0x0ca8) - || (c >= 0x0caa && c <= 0x0cb3) - || (c >= 0x0cb5 && c <= 0x0cb9) - || (c >= 0x0ce0 && c <= 0x0ce1)) - return 1; - - /* Malayalam */ - if ((c >= 0x0d05 && c <= 0x0d0c) - || (c >= 0x0d0e && c <= 0x0d10) - || (c >= 0x0d12 && c <= 0x0d28) - || (c >= 0x0d2a && c <= 0x0d39) - || (c >= 0x0d60 && c <= 0x0d61)) - return 1; - - /* Thai */ - if ((c >= 0x0e01 && c <= 0x0e30) - || (c >= 0x0e32 && c <= 0x0e33) - || (c >= 0x0e40 && c <= 0x0e46) - || (c >= 0x0e4f && c <= 0x0e5b)) - return 1; - - /* Lao */ - if ((c >= 0x0e81 && c <= 0x0e82) - || (c == 0x0e84) - || (c == 0x0e87) - || (c == 0x0e88) - || (c == 0x0e8a) - || (c == 0x0e0d) - || (c >= 0x0e94 && c <= 0x0e97) - || (c >= 0x0e99 && c <= 0x0e9f) - || (c >= 0x0ea1 && c <= 0x0ea3) - || (c == 0x0ea5) - || (c == 0x0ea7) - || (c == 0x0eaa) - || (c == 0x0eab) - || (c >= 0x0ead && c <= 0x0eb0) - || (c == 0x0eb2) - || (c == 0x0eb3) - || (c == 0x0ebd) - || (c >= 0x0ec0 && c <= 0x0ec4) - || (c == 0x0ec6)) - return 1; - - /* Georgian */ - if ((c >= 0x10a0 && c <= 0x10c5) - || (c >= 0x10d0 && c <= 0x10f6)) - return 1; - - /* Hiragana */ - if ((c >= 0x3041 && c <= 0x3094) - || (c >= 0x309b && c <= 0x309e)) - return 1; - - /* Katakana */ - if ((c >= 0x30a1 && c <= 0x30fe)) - return 1; - - /* Bopmofo */ - if ((c >= 0x3105 && c <= 0x312c)) - return 1; - - /* Hangul */ - if ((c >= 0x1100 && c <= 0x1159) - || (c >= 0x1161 && c <= 0x11a2) - || (c >= 0x11a8 && c <= 0x11f9)) - return 1; - - /* CJK Unified Ideographs */ - if ((c >= 0xf900 && c <= 0xfa2d) - || (c >= 0xfb1f && c <= 0xfb36) - || (c >= 0xfb38 && c <= 0xfb3c) - || (c == 0xfb3e) - || (c >= 0xfb40 && c <= 0xfb41) - || (c >= 0xfb42 && c <= 0xfb44) - || (c >= 0xfb46 && c <= 0xfbb1) - || (c >= 0xfbd3 && c <= 0xfd3f) - || (c >= 0xfd50 && c <= 0xfd8f) - || (c >= 0xfd92 && c <= 0xfdc7) - || (c >= 0xfdf0 && c <= 0xfdfb) - || (c >= 0xfe70 && c <= 0xfe72) - || (c == 0xfe74) - || (c >= 0xfe76 && c <= 0xfefc) - || (c >= 0xff21 && c <= 0xff3a) - || (c >= 0xff41 && c <= 0xff5a) - || (c >= 0xff66 && c <= 0xffbe) - || (c >= 0xffc2 && c <= 0xffc7) - || (c >= 0xffca && c <= 0xffcf) - || (c >= 0xffd2 && c <= 0xffd7) - || (c >= 0xffda && c <= 0xffdc) - || (c >= 0x4e00 && c <= 0x9fa5)) - return 1; - - error ("universal-character-name '\\u%04x' not valid in identifier", c); - return 1; -#endif + (*debug_hooks->undef) (SOURCE_LINE (map, line), + (const char *) NODE_NAME (node)); } + +static inline const cpp_token * +get_nonpadding_token (void) +{ + const cpp_token *tok; + timevar_push (TV_CPP); + do + tok = cpp_get_token (parse_in); + while (tok->type == CPP_PADDING); + timevar_pop (TV_CPP); -/* Add the UTF-8 representation of C to the token_buffer. */ + /* The C++ front end does horrible things with the current line + number. To ensure an accurate line number, we must reset it + every time we advance a token. */ + input_line = src_lineno; -static void -utf8_extend_token (c) - int c; + return tok; +} + +int +c_lex (tree *value) { - int shift, mask; + const cpp_token *tok; + location_t atloc; + static bool no_more_pch; - if (c <= 0x0000007f) - { - extend_token (c); - return; - } - else if (c <= 0x000007ff) - shift = 6, mask = 0xc0; - else if (c <= 0x0000ffff) - shift = 12, mask = 0xe0; - else if (c <= 0x001fffff) - shift = 18, mask = 0xf0; - else if (c <= 0x03ffffff) - shift = 24, mask = 0xf8; - else - shift = 30, mask = 0xfc; + retry: + tok = get_nonpadding_token (); - extend_token (mask | (c >> shift)); - do + retry_after_at: + switch (tok->type) { - shift -= 6; - extend_token ((unsigned char) (0x80 | (c >> shift))); - } - while (shift); -} -#endif - -#if 0 -struct try_type -{ - tree *node_var; - char unsigned_flag; - char long_flag; - char long_long_flag; -}; + case CPP_NAME: + *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node)); + break; -struct try_type type_sequence[] = -{ - { &integer_type_node, 0, 0, 0}, - { &unsigned_type_node, 1, 0, 0}, - { &long_integer_type_node, 0, 1, 0}, - { &long_unsigned_type_node, 1, 1, 0}, - { &long_long_integer_type_node, 0, 1, 1}, - { &long_long_unsigned_type_node, 1, 1, 1} -}; -#endif /* 0 */ - -struct pf_args -{ - /* Input */ - const char *str; - int fflag; - int lflag; - int base; - /* Output */ - int conversion_errno; - REAL_VALUE_TYPE value; - tree type; -}; - -static void -parse_float (data) - PTR data; -{ - struct pf_args * args = (struct pf_args *) data; - const char *typename; + case CPP_NUMBER: + { + unsigned int flags = cpp_classify_number (parse_in, tok); - args->conversion_errno = 0; - args->type = double_type_node; - typename = "double"; + switch (flags & CPP_N_CATEGORY) + { + case CPP_N_INVALID: + /* cpplib has issued an error. */ + *value = error_mark_node; + break; - /* The second argument, machine_mode, of REAL_VALUE_ATOF - tells the desired precision of the binary result - of decimal-to-binary conversion. */ + case CPP_N_INTEGER: + *value = interpret_integer (tok, flags); + break; - if (args->fflag) - { - if (args->lflag) - error ("both 'f' and 'l' suffixes on floating constant"); + case CPP_N_FLOATING: + *value = interpret_float (tok, flags); + break; - args->type = float_type_node; - typename = "float"; - } - else if (args->lflag) - { - args->type = long_double_type_node; - typename = "long double"; - } - else if (flag_single_precision_constant) - { - args->type = float_type_node; - typename = "float"; - } + default: + abort (); + } + } + break; - errno = 0; - if (args->base == 16) - args->value = REAL_VALUE_HTOF (args->str, TYPE_MODE (args->type)); - else - args->value = REAL_VALUE_ATOF (args->str, TYPE_MODE (args->type)); - - args->conversion_errno = errno; - /* A diagnostic is required here by some ISO C testsuites. - This is not pedwarn, because some people don't want - an error for this. */ - if (REAL_VALUE_ISINF (args->value) && pedantic) - warning ("floating point number exceeds range of '%s'", typename); -} - -int -c_lex (value) - tree *value; -{ - cpp_token tok; - enum cpp_ttype type; + case CPP_ATSIGN: + /* An @ may give the next token special significance in Objective-C. */ + atloc = input_location; + tok = get_nonpadding_token (); + if (c_dialect_objc ()) + { + tree val; + switch (tok->type) + { + case CPP_NAME: + val = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node)); + if (C_IS_RESERVED_WORD (val) + && OBJC_IS_AT_KEYWORD (C_RID_CODE (val))) + { + *value = val; + return CPP_AT_NAME; + } + break; - retry: - timevar_push (TV_CPP); - cpp_get_token (parse_in, &tok); - timevar_pop (TV_CPP); + case CPP_STRING: + case CPP_WSTRING: + return lex_string (tok, value, true); - /* The C++ front end does horrible things with the current line - number. To ensure an accurate line number, we must reset it - every time we return a token. */ - lex_lineno = cpp_get_line (parse_in)->line; + default: break; + } + } - *value = NULL_TREE; - lineno = lex_lineno; - type = tok.type; - switch (type) - { - case CPP_OPEN_BRACE: indent_level++; break; - case CPP_CLOSE_BRACE: indent_level--; break; + /* ... or not. */ + error ("%Hstray '@' in program", &atloc); + goto retry_after_at; - /* Issue this error here, where we can get at tok.val.c. */ case CPP_OTHER: - if (ISGRAPH (tok.val.c)) - error ("stray '%c' in program", tok.val.c); - else - error ("stray '\\%#o' in program", tok.val.c); + { + cppchar_t c = tok->val.str.text[0]; + + if (c == '"' || c == '\'') + error ("missing terminating %c character", (int) c); + else if (ISGRAPH (c)) + error ("stray '%c' in program", (int) c); + else + error ("stray '\\%o' in program", (int) c); + } goto retry; - - case CPP_NAME: - *value = get_identifier ((const char *)tok.val.node->name); - break; - - case CPP_INT: - case CPP_FLOAT: - case CPP_NUMBER: - *value = lex_number ((const char *)tok.val.str.text, tok.val.str.len); - break; case CPP_CHAR: case CPP_WCHAR: - *value = lex_charconst ((const char *)tok.val.str.text, - tok.val.str.len, tok.type == CPP_WCHAR); + *value = lex_charconst (tok); break; case CPP_STRING: case CPP_WSTRING: - case CPP_OSTRING: - *value = lex_string ((const char *)tok.val.str.text, - tok.val.str.len, tok.type == CPP_WSTRING); + return lex_string (tok, value, false); break; /* These tokens should not be visible outside cpplib. */ @@ -1001,693 +426,335 @@ c_lex (value) case CPP_MACRO_ARG: abort (); - default: break; - } - - return type; -} - -#define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0) - -static tree -lex_number (str, len) - const char *str; - unsigned int len; -{ - int base = 10; - int count = 0; - int largest_digit = 0; - int numdigits = 0; - int overflow = 0; - int c; - tree value; - const char *p; - enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT; - - /* We actually store only HOST_BITS_PER_CHAR bits in each part. - The code below which fills the parts array assumes that a host - int is at least twice as wide as a host char, and that - HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR. - Two HOST_WIDE_INTs is the largest int literal we can store. - In order to detect overflow below, the number of parts (TOTAL_PARTS) - must be exactly the number of parts needed to hold the bits - of two HOST_WIDE_INTs. */ -#define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2) - unsigned int parts[TOTAL_PARTS]; - - /* Optimize for most frequent case. */ - if (len == 1) - { - if (*str == '0') - return integer_zero_node; - else if (*str == '1') - return integer_one_node; - else - return build_int_2 (*str - '0', 0); + default: + *value = NULL_TREE; + break; } - for (count = 0; count < TOTAL_PARTS; count++) - parts[count] = 0; - - /* len is known to be >1 at this point. */ - p = str; - - if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X')) - { - base = 16; - p = str + 2; - } - /* The ISDIGIT check is so we are not confused by a suffix on 0. */ - else if (str[0] == '0' && ISDIGIT (str[1])) + if (! no_more_pch) { - base = 8; - p = str + 1; + no_more_pch = true; + c_common_no_more_pch (); } - do - { - c = *p++; - - if (c == '.') - { - if (base == 16 && pedantic && !flag_isoc99) - pedwarn ("floating constant may not be in radix 16"); - else if (floatflag == AFTER_POINT) - ERROR ("too many decimal points in floating constant"); - else if (floatflag == AFTER_EXPON) - ERROR ("decimal point in exponent - impossible!"); - else - floatflag = AFTER_POINT; - - if (base == 8) - base = 10; - } - else if (c == '_') - /* Possible future extension: silently ignore _ in numbers, - permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000 - but somewhat easier to read. Ada has this? */ - ERROR ("underscore in number"); - else - { - int n; - /* It is not a decimal point. - It should be a digit (perhaps a hex digit). */ - - if (ISDIGIT (c)) - { - n = c - '0'; - } - else if (base <= 10 && (c == 'e' || c == 'E')) - { - base = 10; - floatflag = AFTER_EXPON; - break; - } - else if (base == 16 && (c == 'p' || c == 'P')) - { - floatflag = AFTER_EXPON; - break; /* start of exponent */ - } - else if (base == 16 && c >= 'a' && c <= 'f') - { - n = c - 'a' + 10; - } - else if (base == 16 && c >= 'A' && c <= 'F') - { - n = c - 'A' + 10; - } - else - { - p--; - break; /* start of suffix */ - } - - if (n >= largest_digit) - largest_digit = n; - numdigits++; - - for (count = 0; count < TOTAL_PARTS; count++) - { - parts[count] *= base; - if (count) - { - parts[count] - += (parts[count-1] >> HOST_BITS_PER_CHAR); - parts[count-1] - &= (1 << HOST_BITS_PER_CHAR) - 1; - } - else - parts[0] += n; - } - - /* If the highest-order part overflows (gets larger than - a host char will hold) then the whole number has - overflowed. Record this and truncate the highest-order - part. */ - if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR) - { - overflow = 1; - parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1; - } - } - } - while (p < str + len); - - /* This can happen on input like `int i = 0x;' */ - if (numdigits == 0) - ERROR ("numeric constant with no digits"); + return tok->type; +} - if (largest_digit >= base) - ERROR ("numeric constant contains digits beyond the radix"); +/* Returns the narrowest C-visible unsigned type, starting with the + minimum specified by FLAGS, that can fit VALUE, or itk_none if + there isn't one. */ +static enum integer_type_kind +narrowest_unsigned_type (tree value, unsigned int flags) +{ + enum integer_type_kind itk; - if (floatflag != NOT_FLOAT) - { - tree type; - int imag, fflag, lflag, conversion_errno; - REAL_VALUE_TYPE real; - struct pf_args args; - char *copy; - - if (base == 16 && floatflag != AFTER_EXPON) - ERROR ("hexadecimal floating constant has no exponent"); - - /* Read explicit exponent if any, and put it in tokenbuf. */ - if ((base == 10 && ((c == 'e') || (c == 'E'))) - || (base == 16 && (c == 'p' || c == 'P'))) - { - if (p < str + len) - c = *p++; - if (p < str + len && (c == '+' || c == '-')) - c = *p++; - /* Exponent is decimal, even if string is a hex float. */ - if (! ISDIGIT (c)) - ERROR ("floating constant exponent has no digits"); - while (p < str + len && ISDIGIT (c)) - c = *p++; - if (! ISDIGIT (c)) - p--; - } + if ((flags & CPP_N_WIDTH) == CPP_N_SMALL) + itk = itk_unsigned_int; + else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM) + itk = itk_unsigned_long; + else + itk = itk_unsigned_long_long; - /* Copy the float constant now; we don't want any suffixes in the - string passed to parse_float. */ - copy = alloca (p - str + 1); - memcpy (copy, str, p - str); - copy[p - str] = '\0'; + /* int_fits_type_p must think the type of its first argument is + wider than its second argument, or it won't do the proper check. */ + TREE_TYPE (value) = widest_unsigned_literal_type_node; - /* Now parse suffixes. */ - fflag = lflag = imag = 0; - while (p < str + len) - switch (*p++) - { - case 'f': case 'F': - if (fflag) - ERROR ("more than one 'f' suffix on floating constant"); - else if (warn_traditional && !in_system_header) - warning ("traditional C rejects the 'f' suffix"); + for (; itk < itk_none; itk += 2 /* skip unsigned types */) + if (int_fits_type_p (value, integer_types[itk])) + return itk; - fflag = 1; - break; - - case 'l': case 'L': - if (lflag) - ERROR ("more than one 'l' suffix on floating constant"); - else if (warn_traditional && !in_system_header) - warning ("traditional C rejects the 'l' suffix"); + return itk_none; +} - lflag = 1; - break; +/* Ditto, but narrowest signed type. */ +static enum integer_type_kind +narrowest_signed_type (tree value, unsigned int flags) +{ + enum integer_type_kind itk; - case 'i': case 'I': - case 'j': case 'J': - if (imag) - ERROR ("more than one 'i' or 'j' suffix on floating constant"); - else if (pedantic) - pedwarn ("ISO C forbids imaginary numeric constants"); - imag = 1; - break; + if ((flags & CPP_N_WIDTH) == CPP_N_SMALL) + itk = itk_int; + else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM) + itk = itk_long; + else + itk = itk_long_long; - default: - ERROR ("invalid suffix on floating constant"); - } + /* int_fits_type_p must think the type of its first argument is + wider than its second argument, or it won't do the proper check. */ + TREE_TYPE (value) = widest_unsigned_literal_type_node; - /* Setup input for parse_float() */ - args.str = copy; - args.fflag = fflag; - args.lflag = lflag; - args.base = base; + for (; itk < itk_none; itk += 2 /* skip signed types */) + if (int_fits_type_p (value, integer_types[itk])) + return itk; - /* Convert string to a double, checking for overflow. */ - if (do_float_handler (parse_float, (PTR) &args)) - { - /* Receive output from parse_float() */ - real = args.value; - } - else - /* We got an exception from parse_float() */ - ERROR ("floating constant out of range"); - - /* Receive output from parse_float() */ - conversion_errno = args.conversion_errno; - type = args.type; - -#ifdef ERANGE - /* ERANGE is also reported for underflow, - so test the value to distinguish overflow from that. */ - if (conversion_errno == ERANGE && !flag_traditional && pedantic - && (REAL_VALUES_LESS (dconst1, real) - || REAL_VALUES_LESS (real, dconstm1))) - warning ("floating point number exceeds range of 'double'"); -#endif + return itk_none; +} - /* Create a node with determined type and value. */ - if (imag) - value = build_complex (NULL_TREE, convert (type, integer_zero_node), - build_real (type, real)); - else - value = build_real (type, real); - } +/* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */ +static tree +interpret_integer (const cpp_token *token, unsigned int flags) +{ + tree value, type; + enum integer_type_kind itk; + cpp_num integer; + cpp_options *options = cpp_get_options (parse_in); + + integer = cpp_interpret_integer (parse_in, token, flags); + integer = cpp_num_sign_extend (integer, options->precision); + value = build_int_2_wide (integer.low, integer.high); + + /* The type of a constant with a U suffix is straightforward. */ + if (flags & CPP_N_UNSIGNED) + itk = narrowest_unsigned_type (value, flags); else { - tree trad_type, ansi_type, type; - HOST_WIDE_INT high, low; - int spec_unsigned = 0; - int spec_long = 0; - int spec_long_long = 0; - int spec_imag = 0; - int suffix_lu = 0; - int warn = 0, i; - - trad_type = ansi_type = type = NULL_TREE; - while (p < str + len) + /* The type of a potentially-signed integer constant varies + depending on the base it's in, the standard in use, and the + length suffixes. */ + enum integer_type_kind itk_u = narrowest_unsigned_type (value, flags); + enum integer_type_kind itk_s = narrowest_signed_type (value, flags); + + /* In both C89 and C99, octal and hex constants may be signed or + unsigned, whichever fits tighter. We do not warn about this + choice differing from the traditional choice, as the constant + is probably a bit pattern and either way will work. */ + if ((flags & CPP_N_RADIX) != CPP_N_DECIMAL) + itk = MIN (itk_u, itk_s); + else { - c = *p++; - switch (c) + /* In C99, decimal constants are always signed. + In C89, decimal constants that don't fit in long have + undefined behavior; we try to make them unsigned long. + In GCC's extended C89, that last is true of decimal + constants that don't fit in long long, too. */ + + itk = itk_s; + if (itk_s > itk_u && itk_s > itk_long) { - case 'u': case 'U': - if (spec_unsigned) - error ("two 'u' suffixes on integer constant"); - else if (warn_traditional && !in_system_header) - warning ("traditional C rejects the 'u' suffix"); - - spec_unsigned = 1; - if (spec_long) - suffix_lu = 1; - break; - - case 'l': case 'L': - if (spec_long) + if (!flag_isoc99) { - if (spec_long_long) - error ("three 'l' suffixes on integer constant"); - else if (suffix_lu) - error ("'lul' is not a valid integer suffix"); - else if (c != spec_long) - error ("'Ll' and 'lL' are not valid integer suffixes"); - else if (pedantic && ! flag_isoc99 - && ! in_system_header && warn_long_long) - pedwarn ("ISO C89 forbids long long integer constants"); - spec_long_long = 1; + if (itk_u < itk_unsigned_long) + itk_u = itk_unsigned_long; + itk = itk_u; + warning ("this decimal constant is unsigned only in ISO C90"); } - spec_long = c; - break; - - case 'i': case 'I': case 'j': case 'J': - if (spec_imag) - error ("more than one 'i' or 'j' suffix on integer constant"); - else if (pedantic) - pedwarn ("ISO C forbids imaginary numeric constants"); - spec_imag = 1; - break; - - default: - ERROR ("invalid suffix on integer constant"); + else if (warn_traditional) + warning ("this decimal constant would be unsigned in ISO C90"); } } + } - /* If the literal overflowed, pedwarn about it now. */ - if (overflow) - { - warn = 1; - pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2); - } - - /* This is simplified by the fact that our constant - is always positive. */ - - high = low = 0; - - for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++) - { - high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT - / HOST_BITS_PER_CHAR)] - << (i * HOST_BITS_PER_CHAR)); - low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR); - } - - value = build_int_2 (low, high); - TREE_TYPE (value) = long_long_unsigned_type_node; - - /* If warn_traditional, calculate both the ISO type and the - traditional type, then see if they disagree. - Otherwise, calculate only the type for the dialect in use. */ - if (warn_traditional || flag_traditional) - { - /* Calculate the traditional type. */ - /* Traditionally, any constant is signed; but if unsigned is - specified explicitly, obey that. Use the smallest size - with the right number of bits, except for one special - case with decimal constants. */ - if (! spec_long && base != 10 - && int_fits_type_p (value, unsigned_type_node)) - trad_type = spec_unsigned ? unsigned_type_node : integer_type_node; - /* A decimal constant must be long if it does not fit in - type int. I think this is independent of whether the - constant is signed. */ - else if (! spec_long && base == 10 - && int_fits_type_p (value, integer_type_node)) - trad_type = spec_unsigned ? unsigned_type_node : integer_type_node; - else if (! spec_long_long) - trad_type = (spec_unsigned - ? long_unsigned_type_node - : long_integer_type_node); - else if (int_fits_type_p (value, - spec_unsigned - ? long_long_unsigned_type_node - : long_long_integer_type_node)) - trad_type = (spec_unsigned - ? long_long_unsigned_type_node - : long_long_integer_type_node); - else - trad_type = (spec_unsigned - ? widest_unsigned_literal_type_node - : widest_integer_literal_type_node); - } - if (warn_traditional || ! flag_traditional) - { - /* Calculate the ISO type. */ - if (! spec_long && ! spec_unsigned - && int_fits_type_p (value, integer_type_node)) - ansi_type = integer_type_node; - else if (! spec_long && (base != 10 || spec_unsigned) - && int_fits_type_p (value, unsigned_type_node)) - ansi_type = unsigned_type_node; - else if (! spec_unsigned && !spec_long_long - && int_fits_type_p (value, long_integer_type_node)) - ansi_type = long_integer_type_node; - else if (! spec_long_long - && int_fits_type_p (value, long_unsigned_type_node)) - ansi_type = long_unsigned_type_node; - else if (! spec_unsigned - && int_fits_type_p (value, long_long_integer_type_node)) - ansi_type = long_long_integer_type_node; - else if (int_fits_type_p (value, long_long_unsigned_type_node)) - ansi_type = long_long_unsigned_type_node; - else if (! spec_unsigned - && int_fits_type_p (value, widest_integer_literal_type_node)) - ansi_type = widest_integer_literal_type_node; - else - ansi_type = widest_unsigned_literal_type_node; - } - - type = flag_traditional ? trad_type : ansi_type; - - /* We assume that constants specified in a non-decimal - base are bit patterns, and that the programmer really - meant what they wrote. */ - if (warn_traditional && !in_system_header - && base == 10 && trad_type != ansi_type) - { - if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (ansi_type)) - warning ("width of integer constant changes with -traditional"); - else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (ansi_type)) - warning ("integer constant is unsigned in ISO C, signed with -traditional"); - else - warning ("width of integer constant may change on other systems with -traditional"); - } - - if (pedantic && !flag_traditional && (flag_isoc99 || !spec_long_long) - && !warn - && ((flag_isoc99 - ? TYPE_PRECISION (long_long_integer_type_node) - : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type))) - { - warn = 1; - pedwarn ("integer constant larger than the maximum value of %s", - (flag_isoc99 - ? (TREE_UNSIGNED (type) - ? "an unsigned long long int" - : "a long long int") - : "an unsigned long int")); - } - - if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type)) - warning ("decimal constant is so large that it is unsigned"); - - if (spec_imag) - { - if (TYPE_PRECISION (type) - <= TYPE_PRECISION (integer_type_node)) - value = build_complex (NULL_TREE, integer_zero_node, - convert (integer_type_node, value)); - else - ERROR ("complex integer constant is too wide for 'complex int'"); - } - else if (flag_traditional && !int_fits_type_p (value, type)) - /* The traditional constant 0x80000000 is signed - but doesn't fit in the range of int. - This will change it to -0x80000000, which does fit. */ - { - TREE_TYPE (value) = unsigned_type (type); - value = convert (type, value); - TREE_OVERFLOW (value) = TREE_CONSTANT_OVERFLOW (value) = 0; - } - else - TREE_TYPE (value) = type; + if (itk == itk_none) + /* cpplib has already issued a warning for overflow. */ + type = ((flags & CPP_N_UNSIGNED) + ? widest_unsigned_literal_type_node + : widest_integer_literal_type_node); + else + type = integer_types[itk]; - /* If it's still an integer (not a complex), and it doesn't - fit in the type we choose for it, then pedwarn. */ + if (itk > itk_unsigned_long + && (flags & CPP_N_WIDTH) != CPP_N_LARGE + && ! in_system_header && ! flag_isoc99) + pedwarn ("integer constant is too large for \"%s\" type", + (flags & CPP_N_UNSIGNED) ? "unsigned long" : "long"); - if (! warn - && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE - && ! int_fits_type_p (value, TREE_TYPE (value))) - pedwarn ("integer constant is larger than the maximum value for its type"); - } + TREE_TYPE (value) = type; - if (p < str + len) - error ("missing white space after number '%.*s'", (int) (p - str), str); + /* Convert imaginary to a complex type. */ + if (flags & CPP_N_IMAGINARY) + value = build_complex (NULL_TREE, convert (type, integer_zero_node), value); return value; - - syntax_error: - return integer_zero_node; } +/* Interpret TOKEN, a floating point number with FLAGS as classified + by cpplib. */ static tree -lex_string (str, len, wide) - const char *str; - unsigned int len; - int wide; +interpret_float (const cpp_token *token, unsigned int flags) { + tree type; tree value; - char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1)); - char *q = buf; - const char *p = str, *limit = str + len; - unsigned int c; - unsigned width = wide ? WCHAR_TYPE_SIZE - : TYPE_PRECISION (char_type_node); - -#ifdef MULTIBYTE_CHARS - /* Reset multibyte conversion state. */ - (void) local_mbtowc (NULL_PTR, NULL_PTR, 0); -#endif + REAL_VALUE_TYPE real; + char *copy; + size_t copylen; + const char *typename; - while (p < limit) + /* FIXME: make %T work in error/warning, then we don't need typename. */ + if ((flags & CPP_N_WIDTH) == CPP_N_LARGE) { -#ifdef MULTIBYTE_CHARS - wchar_t wc; - int char_len; - - char_len = local_mbtowc (&wc, p, limit - p); - if (char_len == -1) - { - warning ("Ignoring invalid multibyte character"); - char_len = 1; - c = *p++; - } - else - { - p += char_len; - c = wc; - } -#else - c = *p++; -#endif - - if (c == '\\' && !ignore_escape_flag) - { - p = readescape (p, limit, &c); - if (width < HOST_BITS_PER_INT - && (unsigned) c >= ((unsigned)1 << width)) - pedwarn ("escape sequence out of range for character"); - } - - /* Add this single character into the buffer either as a wchar_t - or as a single byte. */ - if (wide) - { - unsigned charwidth = TYPE_PRECISION (char_type_node); - unsigned bytemask = (1 << charwidth) - 1; - int byte; - - for (byte = 0; byte < WCHAR_BYTES; ++byte) - { - int n; - if (byte >= (int) sizeof (c)) - n = 0; - else - n = (c >> (byte * charwidth)) & bytemask; - if (BYTES_BIG_ENDIAN) - q[WCHAR_BYTES - byte - 1] = n; - else - q[byte] = n; - } - q += WCHAR_BYTES; - } - else - { - *q++ = c; - } + type = long_double_type_node; + typename = "long double"; } - - /* Terminate the string value, either with a single byte zero - or with a wide zero. */ - - if (wide) + else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL + || flag_single_precision_constant) { - memset (q, 0, WCHAR_BYTES); - q += WCHAR_BYTES; + type = float_type_node; + typename = "float"; } else { - *q++ = '\0'; + type = double_type_node; + typename = "double"; } - value = build_string (q - buf, buf); + /* Copy the constant to a nul-terminated buffer. If the constant + has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF + can't handle them. */ + copylen = token->val.str.len; + if ((flags & CPP_N_WIDTH) != CPP_N_MEDIUM) + /* Must be an F or L suffix. */ + copylen--; + if (flags & CPP_N_IMAGINARY) + /* I or J suffix. */ + copylen--; + + copy = alloca (copylen + 1); + memcpy (copy, token->val.str.text, copylen); + copy[copylen] = '\0'; + + real_from_string (&real, copy); + real_convert (&real, TYPE_MODE (type), &real); + + /* A diagnostic is required for "soft" overflow by some ISO C + testsuites. This is not pedwarn, because some people don't want + an error for this. + ??? That's a dubious reason... is this a mandatory diagnostic or + isn't it? -- zw, 2001-08-21. */ + if (REAL_VALUE_ISINF (real) && pedantic) + warning ("floating constant exceeds range of \"%s\"", typename); + + /* Create a node with determined type and value. */ + value = build_real (type, real); + if (flags & CPP_N_IMAGINARY) + value = build_complex (NULL_TREE, convert (type, integer_zero_node), value); - if (wide) - TREE_TYPE (value) = wchar_array_type_node; - else - TREE_TYPE (value) = char_array_type_node; return value; } -static tree -lex_charconst (str, len, wide) - const char *str; - unsigned int len; - int wide; +/* Convert a series of STRING and/or WSTRING tokens into a tree, + performing string constant concatenation. TOK is the first of + these. VALP is the location to write the string into. OBJC_STRING + indicates whether an '@' token preceded the incoming token. + Returns the CPP token type of the result (CPP_STRING, CPP_WSTRING, + or CPP_OBJC_STRING). + + This is unfortunately more work than it should be. If any of the + strings in the series has an L prefix, the result is a wide string + (6.4.5p4). Whether or not the result is a wide string affects the + meaning of octal and hexadecimal escapes (6.4.4.4p6,9). But escape + sequences do not continue across the boundary between two strings in + a series (6.4.5p7), so we must not lose the boundaries. Therefore + cpp_interpret_string takes a vector of cpp_string structures, which + we must arrange to provide. */ + +static enum cpp_ttype +lex_string (const cpp_token *tok, tree *valp, bool objc_string) { - const char *limit = str + len; - int result = 0; - int num_chars = 0; - int chars_seen = 0; - unsigned width = TYPE_PRECISION (char_type_node); - int max_chars; - unsigned int c; tree value; + bool wide = false; + size_t count = 1; + struct obstack str_ob; + cpp_string istr; -#ifdef MULTIBYTE_CHARS - int longest_char = local_mb_cur_max (); - (void) local_mbtowc (NULL_PTR, NULL_PTR, 0); -#endif + /* Try to avoid the overhead of creating and destroying an obstack + for the common case of just one string. */ + cpp_string str = tok->val.str; + cpp_string *strs = &str; - max_chars = TYPE_PRECISION (integer_type_node) / width; - if (wide) - width = WCHAR_TYPE_SIZE; + if (tok->type == CPP_WSTRING) + wide = true; - while (str < limit) + tok = get_nonpadding_token (); + if (c_dialect_objc () && tok->type == CPP_ATSIGN) { -#ifdef MULTIBYTE_CHARS - wchar_t wc; - int char_len; - - char_len = local_mbtowc (&wc, str, limit - str); - if (char_len == -1) - { - warning ("Ignoring invalid multibyte character"); - char_len = 1; - c = *str++; - } - else - { - p += char_len; - c = wc; - } -#else - c = *str++; -#endif + objc_string = true; + tok = get_nonpadding_token (); + } + if (tok->type == CPP_STRING || tok->type == CPP_WSTRING) + { + gcc_obstack_init (&str_ob); + obstack_grow (&str_ob, &str, sizeof (cpp_string)); - ++chars_seen; - if (c == '\\') + do { - str = readescape (str, limit, &c); - 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 - - /* Merge character into result; ignore excess chars. */ - num_chars += (width / TYPE_PRECISION (char_type_node)); - if (num_chars < max_chars + 1) - { - if (width < HOST_BITS_PER_INT) - result = (result << width) | (c & ((1 << width) - 1)); - else - result = c; + count++; + if (tok->type == CPP_WSTRING) + wide = true; + obstack_grow (&str_ob, &tok->val.str, sizeof (cpp_string)); + + tok = get_nonpadding_token (); + if (c_dialect_objc () && tok->type == CPP_ATSIGN) + { + objc_string = true; + tok = get_nonpadding_token (); + } } + while (tok->type == CPP_STRING || tok->type == CPP_WSTRING); + strs = obstack_finish (&str_ob); } - if (chars_seen == 0) - error ("empty character constant"); - else if (num_chars > max_chars) - { - num_chars = max_chars; - error ("character constant too long"); - } - else if (chars_seen != 1 && ! flag_traditional && warn_multichar) - warning ("multi-character character constant"); + /* We have read one more token than we want. */ + _cpp_backup_tokens (parse_in, 1); + + if (count > 1 && !objc_string && warn_traditional && !in_system_header) + warning ("traditional C rejects string constant concatenation"); - /* If char type is signed, sign-extend the constant. */ - if (! wide) + if (cpp_interpret_string (parse_in, strs, count, &istr, wide)) { - int num_bits = num_chars * width; - if (num_bits == 0) - /* We already got an error; avoid invalid shift. */ - value = build_int_2 (0, 0); - else if (TREE_UNSIGNED (char_type_node) - || ((result >> (num_bits - 1)) & 1) == 0) - value = build_int_2 (result & (~(unsigned HOST_WIDE_INT) 0 - >> (HOST_BITS_PER_WIDE_INT - num_bits)), - 0); - else - value = build_int_2 (result | ~(~(unsigned HOST_WIDE_INT) 0 - >> (HOST_BITS_PER_WIDE_INT - num_bits)), - -1); - /* In C, a character constant has type 'int'; in C++, 'char'. */ - if (chars_seen <= 1 && c_language == clk_cplusplus) - TREE_TYPE (value) = char_type_node; - else - TREE_TYPE (value) = integer_type_node; + value = build_string (istr.len, (char *)istr.text); + free ((void *)istr.text); } else { - value = build_int_2 (result, 0); - TREE_TYPE (value) = wchar_type_node; + /* Callers cannot generally handle error_mark_node in this context, + so return the empty string instead. cpp_interpret_string has + issued an error. */ + if (wide) + value = build_string (TYPE_PRECISION (wchar_type_node) + / TYPE_PRECISION (char_type_node), + "\0\0\0"); /* widest supported wchar_t + is 32 bits */ + else + value = build_string (1, ""); } + TREE_TYPE (value) = wide ? wchar_array_type_node : char_array_type_node; + *valp = fix_string_type (value); + + if (strs != &str) + obstack_free (&str_ob, 0); + + return objc_string ? CPP_OBJC_STRING : wide ? CPP_WSTRING : CPP_STRING; +} + +/* Converts a (possibly wide) character constant token into a tree. */ +static tree +lex_charconst (const cpp_token *token) +{ + cppchar_t result; + tree type, value; + unsigned int chars_seen; + int unsignedp; + + result = cpp_interpret_charconst (parse_in, token, + &chars_seen, &unsignedp); + + /* Cast to cppchar_signed_t to get correct sign-extension of RESULT + before possibly widening to HOST_WIDE_INT for build_int_2. */ + if (unsignedp || (cppchar_signed_t) result >= 0) + value = build_int_2 (result, 0); + else + value = build_int_2 ((cppchar_signed_t) result, -1); + + if (token->type == CPP_WCHAR) + type = wchar_type_node; + /* In C, a character constant has type 'int'. + In C++ 'char', but multi-char charconsts have type 'int'. */ + else if (!c_dialect_cxx () || chars_seen > 1) + type = integer_type_node; + else + type = char_type_node; + + TREE_TYPE (value) = type; return value; }