/* Bison Grammar Scanner -*- C -*- Copyright (C) 2002 Free Software Foundation, Inc. This file is part of Bison, the GNU Compiler Compiler. This program 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 of the License, or (at your option) any later version. This program 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 this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ %option debug nodefault noyywrap nounput never-interactive stack %option prefix="gram_" outfile="lex.yy.c" %{ #include "system.h" #include "mbswidth.h" #include "complain.h" #include "quote.h" #include "getargs.h" #include "gram.h" #include "reader.h" /* Each time we match a string, move the end cursor to its end. */ #define YY_USER_INIT \ do { \ LOCATION_RESET (*yylloc); \ yylloc->file = infile; \ /* This is only to avoid GCC warnings. */ \ if (yycontrol) {;}; \ } while (0) #define YY_USER_ACTION extend_location (yylloc, yytext, yyleng); #define YY_STEP LOCATION_STEP (*yylloc) #define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size)) /* Read bytes from FP into buffer BUF of size SIZE. Return the number of bytes read. Remove '\r' from input, treating \r\n and isolated \r as \n. */ static size_t no_cr_read (FILE *fp, char *buf, size_t size) { size_t s = fread (buf, 1, size, fp); if (s) { char *w = memchr (buf, '\r', s); if (w) { char const *r = ++w; char const *lim = buf + s; for (;;) { /* Found an '\r'. Treat it like '\n', but ignore any '\n' that immediately follows. */ w[-1] = '\n'; if (r == lim) { int ch = getc (fp); if (ch != '\n' && ungetc (ch, fp) != ch) break; } else if (*r == '\n') r++; /* Copy until the next '\r'. */ do { if (r == lim) return w - buf; } while ((*w++ = *r++) != '\r'); } return w - buf; } } return s; } /* Extend *LOC to account for token TOKEN of size SIZE. */ static void extend_location (location_t *loc, char const *token, int size) { int line = loc->last_line; int column = loc->last_column; char const *p0 = token; char const *p = token; char const *lim = token + size; for (p = token; p < lim; p++) switch (*p) { case '\r': /* \r shouldn't survive no_cr_read. */ abort (); case '\n': line++; column = 1; p0 = p + 1; break; case '\t': column += mbsnwidth (p0, p - p0, 0); column += 8 - ((column - 1) & 7); p0 = p + 1; break; } loc->last_line = line; loc->last_column = column + mbsnwidth (p0, p - p0, 0); } /* Report an unexpected end of file at LOC. A token or comment began with TOKEN_START, but an end of file was encountered and the expected TOKEN_END was missing. */ static void unexpected_end_of_file (location_t loc, char const *token_start, char const *token_end) { complain_at (loc, _("unexpected end of file in `%s ... %s'"), token_start, token_end); } /* STRING_OBSTACK -- Used to store all the characters that we need to keep (to construct ID, STRINGS etc.). Use the following macros to use it. Use YY_OBS_GROW to append what has just been matched, and YY_OBS_FINISH to end the string (it puts the ending 0). YY_OBS_FINISH also stores this string in LAST_STRING, which can be used, and which is used by YY_OBS_FREE to free the last string. */ static struct obstack string_obstack; char *last_string; #define YY_OBS_GROW \ obstack_grow (&string_obstack, yytext, yyleng) #define YY_OBS_FINISH \ do { \ obstack_1grow (&string_obstack, '\0'); \ last_string = obstack_finish (&string_obstack); \ } while (0) #define YY_OBS_FREE \ do { \ obstack_free (&string_obstack, last_string); \ } while (0) void scanner_last_string_free (void) { YY_OBS_FREE; } static int percent_percent_count = 0; /* Within well-formed rules, RULE_LENGTH is the number of values in the current rule so far, which says where to find `$0' with respect to the top of the stack. It is not the same as the rule->length in the case of mid rule actions. Outside of well-formed rules, RULE_LENGTH has an undefined value. */ static int rule_length; static void handle_dollar (braced_code_t code_kind, char *cp, location_t location); static void handle_at (braced_code_t code_kind, char *cp, location_t location); static void handle_syncline (char *args, location_t *location); static int convert_ucn_to_byte (char const *hex_text); %} %x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT %x SC_STRING SC_CHARACTER %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] id {letter}({letter}|[0-9])* directive %{letter}({letter}|[0-9]|-)* int [0-9]+ /* POSIX says that a tag must be both an id and a C union member, but historically almost any character is allowed in a tag. We disallow NUL and newline, as this simplifies our implementation. */ tag [^\0\n>]+ /* Zero or more instances of backslash-newline. Following GCC, allow white space between the backslash and the newline. */ splice (\\[ \f\t\v]*\n)* %% %{ int braces_level IF_LINT (= 0); /* At each yylex invocation, mark the current position as the start of the next token. */ YY_STEP; %} /*----------------------------. | Scanning Bison directives. | `----------------------------*/ { "%binary" return PERCENT_NONASSOC; "%debug" return PERCENT_DEBUG; "%define" return PERCENT_DEFINE; "%defines" return PERCENT_DEFINES; "%destructor" return PERCENT_DESTRUCTOR; "%dprec" return PERCENT_DPREC; "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE; "%expect" return PERCENT_EXPECT; "%file-prefix" return PERCENT_FILE_PREFIX; "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC; "%glr-parser" return PERCENT_GLR_PARSER; "%left" return PERCENT_LEFT; "%locations" return PERCENT_LOCATIONS; "%merge" return PERCENT_MERGE; "%name"[-_]"prefix" return PERCENT_NAME_PREFIX; "%no"[-_]"lines" return PERCENT_NO_LINES; "%nonassoc" return PERCENT_NONASSOC; "%nterm" return PERCENT_NTERM; "%output" return PERCENT_OUTPUT; "%parse-param" return PERCENT_PARSE_PARAM; "%prec" rule_length--; return PERCENT_PREC; "%printer" return PERCENT_PRINTER; "%pure"[-_]"parser" return PERCENT_PURE_PARSER; "%right" return PERCENT_RIGHT; "%lex-param" return PERCENT_LEX_PARAM; "%skeleton" return PERCENT_SKELETON; "%start" return PERCENT_START; "%term" return PERCENT_TOKEN; "%token" return PERCENT_TOKEN; "%token"[-_]"table" return PERCENT_TOKEN_TABLE; "%type" return PERCENT_TYPE; "%union" return PERCENT_UNION; "%verbose" return PERCENT_VERBOSE; "%yacc" return PERCENT_YACC; {directive} { complain_at (*yylloc, _("invalid directive: %s"), quote (yytext)); YY_STEP; } ^"#line "{int}" \""[^\"]*"\"\n" handle_syncline (yytext + strlen ("#line "), yylloc); YY_STEP; "=" return EQUAL; ":" rule_length = 0; return COLON; "|" rule_length = 0; return PIPE; "," return COMMA; ";" return SEMICOLON; [ \f\n\t\v]+ YY_STEP; {id} { yylval->symbol = symbol_get (yytext, *yylloc); rule_length++; return ID; } {int} { unsigned long num; errno = 0; num = strtoul (yytext, 0, 10); if (INT_MAX < num || errno) { complain_at (*yylloc, _("integer out of range: %s"), quote (yytext)); num = INT_MAX; } yylval->integer = num; return INT; } /* Characters. We don't check there is only one. */ "'" YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER); /* Strings. */ "\"" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING); /* Comments. */ "/*" BEGIN SC_YACC_COMMENT; "//".* YY_STEP; /* Prologue. */ "%{" yy_push_state (SC_PROLOGUE); /* Code in between braces. */ "{" YY_OBS_GROW; braces_level = 0; yy_push_state (SC_BRACED_CODE); /* A type. */ "<"{tag}">" { obstack_grow (&string_obstack, yytext + 1, yyleng - 2); YY_OBS_FINISH; yylval->string = last_string; return TYPE; } "%%" { if (++percent_percent_count == 2) yy_push_state (SC_EPILOGUE); return PERCENT_PERCENT; } . { complain_at (*yylloc, _("invalid character: %s"), quote (yytext)); YY_STEP; } } /*-------------------------------------------------------------------. | Whatever the start condition (but those which correspond to | | entities `swallowed' by Bison: SC_YACC_COMMENT, SC_ESCAPED_STRING, | | and SC_ESCAPED_CHARACTER), no M4 character must escape as is. | `-------------------------------------------------------------------*/ { \[ obstack_sgrow (&string_obstack, "@<:@"); \] obstack_sgrow (&string_obstack, "@:>@"); } /*---------------------------------------------------------------. | Scanning a Yacc comment. The initial `/ *' is already eaten. | `---------------------------------------------------------------*/ { "*/" { YY_STEP; BEGIN INITIAL; } [^*]+|"*" ; <> { unexpected_end_of_file (*yylloc, "/*", "*/"); BEGIN INITIAL; } } /*------------------------------------------------------------. | Scanning a C comment. The initial `/ *' is already eaten. | `------------------------------------------------------------*/ { "*"{splice}"/" YY_OBS_GROW; yy_pop_state (); [^*\[\]]+|"*" YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "/*", "*/"); yy_pop_state (); } } /*--------------------------------------------------------------. | Scanning a line comment. The initial `//' is already eaten. | `--------------------------------------------------------------*/ { "\n" YY_OBS_GROW; yy_pop_state (); ([^\n\[\]]|{splice})+ YY_OBS_GROW; <> yy_pop_state (); } /*----------------------------------------------------------------. | Scanning a C string, including its escapes. The initial `"' is | | already eaten. | `----------------------------------------------------------------*/ { "\"" { assert (yy_top_state () == INITIAL); YY_OBS_GROW; YY_OBS_FINISH; yylval->string = last_string; yy_pop_state (); rule_length++; return STRING; } [^\"\\]+ YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "\"", "\""); assert (yy_top_state () == INITIAL); YY_OBS_FINISH; yylval->string = last_string; yy_pop_state (); return STRING; } } /*---------------------------------------------------------------. | Scanning a C character, decoding its escapes. The initial "'" | | is already eaten. | `---------------------------------------------------------------*/ { "'" { YY_OBS_GROW; assert (yy_top_state () == INITIAL); { YY_OBS_FINISH; yylval->symbol = symbol_get (last_string, *yylloc); symbol_class_set (yylval->symbol, token_sym, *yylloc); symbol_user_token_number_set (yylval->symbol, (unsigned char) last_string[1], *yylloc); YY_OBS_FREE; yy_pop_state (); rule_length++; return ID; } } [^\'\\]+ YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "'", "'"); assert (yy_top_state () == INITIAL); YY_OBS_FINISH; yylval->string = last_string; yy_pop_state (); return CHARACTER; } } /*----------------------------. | Decode escaped characters. | `----------------------------*/ { \\[0-7]{1,3} { unsigned long c = strtoul (yytext + 1, 0, 8); if (UCHAR_MAX < c) { complain_at (*yylloc, _("invalid escape sequence: %s"), quote (yytext)); YY_STEP; } else obstack_1grow (&string_obstack, c); } \\x[0-9a-fA-F]+ { unsigned long c; errno = 0; c = strtoul (yytext + 2, 0, 16); if (UCHAR_MAX < c || errno) { complain_at (*yylloc, _("invalid escape sequence: %s"), quote (yytext)); YY_STEP; } else obstack_1grow (&string_obstack, c); } \\a obstack_1grow (&string_obstack, '\a'); \\b obstack_1grow (&string_obstack, '\b'); \\f obstack_1grow (&string_obstack, '\f'); \\n obstack_1grow (&string_obstack, '\n'); \\r obstack_1grow (&string_obstack, '\r'); \\t obstack_1grow (&string_obstack, '\t'); \\v obstack_1grow (&string_obstack, '\v'); \\[\"\'?\\] obstack_1grow (&string_obstack, yytext[1]); \\(u|U[0-9a-fA-F]{4})[0-9a-fA-F]{4} { int c = convert_ucn_to_byte (yytext); if (c < 0) { complain_at (*yylloc, _("invalid escape sequence: %s"), quote (yytext)); YY_STEP; } else obstack_1grow (&string_obstack, c); } \\(.|\n) { complain_at (*yylloc, _("unrecognized escape sequence: %s"), quote (yytext)); YY_OBS_GROW; } /* FLex wants this rule, in case of a `\<>'. */ \\ YY_OBS_GROW; } /*----------------------------------------------------------. | Scanning a C character without decoding its escapes. The | | initial "'" is already eaten. | `----------------------------------------------------------*/ { "'" { YY_OBS_GROW; assert (yy_top_state () != INITIAL); yy_pop_state (); } [^'\[\]\\]+ YY_OBS_GROW; \\{splice}[^\[\]] YY_OBS_GROW; {splice} YY_OBS_GROW; /* Needed for `\<>', `\\<>[', and `\\<>]'. */ \\ YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "'", "'"); assert (yy_top_state () != INITIAL); yy_pop_state (); } } /*----------------------------------------------------------------. | Scanning a C string, without decoding its escapes. The initial | | `"' is already eaten. | `----------------------------------------------------------------*/ { "\"" { assert (yy_top_state () != INITIAL); YY_OBS_GROW; yy_pop_state (); } [^\"\[\]\\]+ YY_OBS_GROW; \\{splice}[^\[\]] YY_OBS_GROW; {splice} YY_OBS_GROW; /* Needed for `\<>', `\\<>[', and `\\<>]'. */ \\ YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "\"", "\""); assert (yy_top_state () != INITIAL); yy_pop_state (); } } /*---------------------------------------------------. | Strings, comments etc. can be found in user code. | `---------------------------------------------------*/ { /* Characters. We don't check there is only one. */ "'" YY_OBS_GROW; yy_push_state (SC_CHARACTER); /* Strings. */ "\"" YY_OBS_GROW; yy_push_state (SC_STRING); /* Comments. */ "/"{splice}"*" YY_OBS_GROW; yy_push_state (SC_COMMENT); "/"{splice}"/" YY_OBS_GROW; yy_push_state (SC_LINE_COMMENT); /* Not comments. */ "/" YY_OBS_GROW; } /*---------------------------------------------------------------. | Scanning some code in braces (%union and actions). The initial | | "{" is already eaten. | `---------------------------------------------------------------*/ { "{"|"<"{splice}"%" YY_OBS_GROW; braces_level++; "%"{splice}">" YY_OBS_GROW; braces_level--; "}" { YY_OBS_GROW; braces_level--; if (braces_level < 0) { yy_pop_state (); YY_OBS_FINISH; yylval->string = last_string; rule_length++; return BRACED_CODE; } } "$"("<"{tag}">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code, yytext, *yylloc); } "@"(-?[0-9]+|"$") { handle_at (current_braced_code, yytext, *yylloc); } /* `"<"{splice}"<"' tokenizes `<<%' correctly (as `<<' `%') rather than incorrrectly (as `<' `<%'). */ [^\"$%\'/<@\[\]\{\}]+|[$%/<@]|"<"{splice}"<" YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "{", "}"); yy_pop_state (); YY_OBS_FINISH; yylval->string = last_string; return BRACED_CODE; } } /*--------------------------------------------------------------. | Scanning some prologue: from "%{" (already scanned) to "%}". | `--------------------------------------------------------------*/ { "%}" { yy_pop_state (); YY_OBS_FINISH; yylval->string = last_string; return PROLOGUE; } [^%\[\]/\'\"]+ YY_OBS_GROW; "%" YY_OBS_GROW; <> { unexpected_end_of_file (*yylloc, "%{", "%}"); yy_pop_state (); YY_OBS_FINISH; yylval->string = last_string; return PROLOGUE; } } /*---------------------------------------------------------------. | Scanning the epilogue (everything after the second "%%", which | | has already been eaten). | `---------------------------------------------------------------*/ { [^\[\]]+ YY_OBS_GROW; <> { yy_pop_state (); YY_OBS_FINISH; yylval->string = last_string; return EPILOGUE; } } %% /*------------------------------------------------------------------. | TEXT is pointing to a wannabee semantic value (i.e., a `$'). | | | | Possible inputs: $[]($|integer) | | | | Output to the STRING_OBSTACK a reference to this semantic value. | `------------------------------------------------------------------*/ static inline void handle_action_dollar (char *text, location_t location) { const char *type_name = NULL; char *cp = text + 1; /* Get the type name if explicit. */ if (*cp == '<') { type_name = ++cp; while (*cp != '>') ++cp; *cp = '\0'; ++cp; } if (*cp == '$') { if (!type_name) type_name = symbol_list_n_type_name_get (current_rule, location, 0); if (!type_name && typed) complain_at (location, _("$$ of `%s' has no declared type"), current_rule->sym->tag); if (!type_name) type_name = ""; obstack_fgrow1 (&string_obstack, "]b4_lhs_value([%s])[", type_name); } else { long num; errno = 0; num = strtol (cp, 0, 10); if (INT_MIN <= num && num <= rule_length && ! errno) { int n = num; if (!type_name && n > 0) type_name = symbol_list_n_type_name_get (current_rule, location, n); if (!type_name && typed) complain_at (location, _("$%d of `%s' has no declared type"), n, current_rule->sym->tag); if (!type_name) type_name = ""; obstack_fgrow3 (&string_obstack, "]b4_rhs_value([%d], [%d], [%s])[", rule_length, n, type_name); } else complain_at (location, _("integer out of range: %s"), quote (text)); } } /*---------------------------------------------------------------. | TEXT is expected to be $$ in some code associated to a symbol: | | destructor or printer. | `---------------------------------------------------------------*/ static inline void handle_symbol_code_dollar (char *text, location_t location) { char *cp = text + 1; if (*cp == '$') obstack_sgrow (&string_obstack, "]b4_dollar_dollar["); else complain_at (location, _("invalid value: %s"), quote (text)); } /*-----------------------------------------------------------------. | Dispatch onto handle_action_dollar, or handle_destructor_dollar, | | depending upon CODE_KIND. | `-----------------------------------------------------------------*/ static void handle_dollar (braced_code_t braced_code_kind, char *text, location_t location) { switch (braced_code_kind) { case action_braced_code: handle_action_dollar (text, location); break; case destructor_braced_code: case printer_braced_code: handle_symbol_code_dollar (text, location); break; } } /*------------------------------------------------------. | TEXT is a location token (i.e., a `@...'). Output to | | STRING_OBSTACK a reference to this location. | `------------------------------------------------------*/ static inline void handle_action_at (char *text, location_t location) { char *cp = text + 1; locations_flag = 1; if (*cp == '$') { obstack_sgrow (&string_obstack, "]b4_lhs_location["); } else { long num; errno = 0; num = strtol (cp, 0, 10); if (INT_MIN <= num && num <= rule_length && ! errno) { int n = num; obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[", rule_length, n); } else complain_at (location, _("integer out of range: %s"), quote (text)); } } /*---------------------------------------------------------------. | TEXT is expected to be @$ in some code associated to a symbol: | | destructor or printer. | `---------------------------------------------------------------*/ static inline void handle_symbol_code_at (char *text, location_t location) { char *cp = text + 1; if (*cp == '$') obstack_sgrow (&string_obstack, "]b4_at_dollar["); else complain_at (location, _("invalid value: %s"), quote (text)); } /*-------------------------------------------------------------------. | Dispatch onto handle_action_at, or handle_destructor_at, depending | | upon CODE_KIND. | `-------------------------------------------------------------------*/ static void handle_at (braced_code_t braced_code_kind, char *text, location_t location) { switch (braced_code_kind) { case action_braced_code: handle_action_at (text, location); break; case destructor_braced_code: case printer_braced_code: handle_symbol_code_at (text, location); break; } } /*------------------------------------------------------------------. | Convert universal character name UCN to a single-byte character, | | and return that character. Return -1 if UCN does not correspond | | to a single-byte character. | `------------------------------------------------------------------*/ static int convert_ucn_to_byte (char const *ucn) { unsigned long code = strtoul (ucn + 2, 0, 16); /* FIXME: Currently we assume Unicode-compatible unibyte characters on ASCII hosts (i.e., Latin-1 on hosts with 8-bit bytes). On non-ASCII hosts we support only the portable C character set. These limitations should be removed once we add support for multibyte characters. */ if (UCHAR_MAX < code) return -1; #if ! ('$' == 0x24 && '@' == 0x40 && '`' == 0x60 && '~' == 0x7e) { /* A non-ASCII host. Use CODE to index into a table of the C basic execution character set, which is guaranteed to exist on all Standard C platforms. This table also includes '$', '@', and '`', which are not in the basic execution character set but which are unibyte characters on all the platforms that we know about. */ static signed char const table[] = { '\0', -1, -1, -1, -1, -1, -1, '\a', '\b', '\t', '\n', '\v', '\f', '\r', -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, ' ', '!', '"', '#', '$', '%', '&', '\'', '(', ')', '*', '+', ',', '-', '.', '/', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', ':', ';', '<', '=', '>', '?', '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', '{', '|', '}', '~' }; code = code < sizeof table ? table[code] : -1; } #endif return code; } /*----------------------------------------------------------------. | Handle `#line INT "FILE"'. ARGS has already skipped `#line '. | `----------------------------------------------------------------*/ static void handle_syncline (char *args, location_t *location) { int lineno = strtol (args, &args, 10); const char *file = NULL; file = strchr (args, '"') + 1; *strchr (file, '"') = 0; /* FIXME: Leaking... Can't free, as some locations are still pointing to the old file name. */ infile = xstrdup (file); location->file = infile; location->last_line = lineno; } /*-------------------------. | Initialize the scanner. | `-------------------------*/ void scanner_initialize (void) { obstack_init (&string_obstack); } /*-----------------------------------------------. | Free all the memory allocated to the scanner. | `-----------------------------------------------*/ void scanner_free (void) { obstack_free (&string_obstack, 0); /* Reclaim Flex's buffers. */ yy_delete_buffer (YY_CURRENT_BUFFER); }