X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/83adb046bf32ee358a0f699bd049f3a3fd2d936e..782e8187185a009ce698ad294c87e49a0a30c3a1:/src/scan-gram.l diff --git a/src/scan-gram.l b/src/scan-gram.l index a1fbac5e..8e48148c 100644 --- a/src/scan-gram.l +++ b/src/scan-gram.l @@ -1,12 +1,12 @@ /* Bison Grammar Scanner -*- C -*- - Copyright (C) 2002 Free Software Foundation, Inc. + Copyright (C) 2002-2012 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 + 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 + the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, @@ -15,135 +15,145 @@ 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 -*/ + along with this program. If not, see . */ -%option debug nodefault noyywrap never-interactive +%option debug nodefault noinput nounput noyywrap never-interactive %option prefix="gram_" outfile="lex.yy.c" %{ -#include "system.h" - +/* Work around a bug in flex 2.5.31. See Debian bug 333231 + . */ +#undef gram_wrap +#define gram_wrap() 1 + +#define FLEX_PREFIX(Id) gram_ ## Id +#include + +#include +#include +#include +#include +#include +#include + +#include #include -#include #include -#include "complain.h" -#include "files.h" -#include "getargs.h" -#include "gram.h" -#include "reader.h" -#include "uniqstr.h" - -#define YY_USER_INIT \ - do \ - { \ - scanner_cursor.file = current_file; \ - scanner_cursor.line = 1; \ - scanner_cursor.column = 1; \ - } \ - while (0) +#include + +#define YY_DECL GRAM_LEX_DECL + +#define YY_USER_INIT \ + code_start = scanner_cursor = loc->start; \ /* Location of scanner cursor. */ -boundary scanner_cursor; +static boundary scanner_cursor; -static void adjust_location (location *, char const *, size_t); -#define YY_USER_ACTION adjust_location (loc, yytext, yyleng); +#define YY_USER_ACTION location_compute (loc, &scanner_cursor, yytext, yyleng); static size_t no_cr_read (FILE *, char *, size_t); #define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size)) +#define RETURN_PERCENT_PARAM(Value) \ + RETURN_VALUE(PERCENT_PARAM, param, param_ ## Value) -/* OBSTACK_FOR_STRING -- Used to store all the characters that we need to - keep (to construct ID, STRINGS etc.). Use the following macros to - use it. +#define RETURN_PERCENT_FLAG(Value) \ + RETURN_VALUE(PERCENT_FLAG, uniqstr, uniqstr_new (Value)) - Use STRING_GROW to append what has just been matched, and - STRING_FINISH to end the string (it puts the ending 0). - STRING_FINISH also stores this string in LAST_STRING, which can be - used, and which is used by STRING_FREE to free the last string. */ +#define RETURN_VALUE(Token, Field, Value) \ + do { \ + val->Field = Value; \ + return Token; \ + } while (0) -static struct obstack obstack_for_string; +#define ROLLBACK_CURRENT_TOKEN \ + do { \ + scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0); \ + yyless (0); \ + } while (0) /* A string representing the most recently saved token. */ static char *last_string; - -#define STRING_GROW \ - obstack_grow (&obstack_for_string, yytext, yyleng) - -#define STRING_FINISH \ - do { \ - obstack_1grow (&obstack_for_string, '\0'); \ - last_string = obstack_finish (&obstack_for_string); \ - } while (0) - -#define STRING_FREE \ - obstack_free (&obstack_for_string, last_string) +/* Bracketed identifier. */ +static uniqstr bracketed_id_str = 0; +static location bracketed_id_loc; +static boundary bracketed_id_start; +static int bracketed_id_context_state = 0; void -scanner_last_string_free (void) +gram_scanner_last_string_free (void) { STRING_FREE; } -/* 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 (int token_type, char *cp, location loc); -static void handle_at (int token_type, char *cp, location loc); -static void handle_syncline (char *args); +static void handle_syncline (char *, location); +static unsigned long int scan_integer (char const *p, int base, location loc); static int convert_ucn_to_byte (char const *hex_text); -static void unexpected_end_of_file (boundary, char const *); +static void unexpected_eof (boundary, char const *); +static void unexpected_newline (boundary, char const *); %} -%x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT -%x SC_STRING SC_CHARACTER -%x SC_AFTER_IDENTIFIER + /* A C-like comment in directives/rules. */ +%x SC_YACC_COMMENT + /* Strings and characters in directives/rules. */ %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER -%x SC_PRE_CODE SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE + /* A identifier was just read in directives/rules. Special state + to capture the sequence 'identifier :'. */ +%x SC_AFTER_IDENTIFIER + /* A complex tag, with nested angles brackets. */ +%x SC_TAG + + /* Four types of user code: + - prologue (code between '%{' '%}' in the first section, before %%); + - actions, printers, union, etc, (between braced in the middle section); + - epilogue (everything after the second %%). + - predicate (code between '%?{' and '{' in middle section); */ +%x SC_PROLOGUE SC_BRACED_CODE SC_EPILOGUE SC_PREDICATE + /* C and C++ comments in code. */ +%x SC_COMMENT SC_LINE_COMMENT + /* Strings and characters in code. */ +%x SC_STRING SC_CHARACTER + /* Bracketed identifiers support. */ +%x SC_BRACKETED_ID SC_RETURN_BRACKETED_ID -letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] -id {letter}({letter}|[0-9])* -directive %{letter}({letter}|[0-9]|-)* -int [0-9]+ +letter [.abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_] +id {letter}({letter}|[-0-9])* +directive %{id} +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>]+ + NUL, as this simplifies our implementation. We disallow angle + bracket to match them in nested pairs: several languages use them + for generics/template types. */ +tag [^\0<>]+ /* Zero or more instances of backslash-newline. Following GCC, allow white space between the backslash and the newline. */ -splice (\\[ \f\t\v]*\n)* +splice (\\[ \f\t\v]*\n)* %% %{ - /* Nesting level of the current code in braces. */ - int braces_level IF_LINT (= 0); + /* Nesting level. Either for nested braces, or nested angle brackets + (but not mixed). */ + int nesting PACIFY_CC (= 0); /* Parent context state, when applicable. */ - int context_state IF_LINT (= 0); - - /* Token type to return, when applicable. */ - int token_type IF_LINT (= 0); + int context_state PACIFY_CC (= 0); /* Location of most recent identifier, when applicable. */ - location id_loc IF_LINT (= *loc); + location id_loc PACIFY_CC (= empty_location); - /* Where containing code started, when applicable. */ - boundary code_start IF_LINT (= loc->start); + /* Where containing code started, when applicable. Its initial + value is relevant only when yylex is invoked in the SC_EPILOGUE + start condition. */ + boundary code_start = scanner_cursor; /* Where containing comment or string or character literal started, when applicable. */ - boundary token_start IF_LINT (= loc->start); + boundary token_start PACIFY_CC (= scanner_cursor); %} @@ -151,12 +161,13 @@ splice (\\[ \f\t\v]*\n)* | Scanning white space. | `-----------------------*/ - + { - [ \f\n\t\v] ; - "," warn_at (*loc, _("stray `,' treated as white space")); - - /* Comments. */ + /* Comments and white space. */ + "," { + complain (loc, Wother, _("stray ',' treated as white space")); + } + [ \f\n\t\v] | "//".* ; "/*" { token_start = loc->start; @@ -167,7 +178,7 @@ splice (\\[ \f\t\v]*\n)* /* #line directives are not documented, and may be withdrawn or modified in future versions of Bison. */ ^"#line "{int}" \"".*"\"\n" { - handle_syncline (yytext + sizeof "#line " - 1); + handle_syncline (yytext + sizeof "#line " - 1, *loc); } } @@ -175,76 +186,95 @@ splice (\\[ \f\t\v]*\n)* /*----------------------------. | Scanning Bison directives. | `----------------------------*/ + + /* For directives that are also command line options, the regex must be + "%..." + after "[-_]"s are removed, and the directive must match the --long + option name, with a single string argument. Otherwise, add exceptions + to ../build-aux/cross-options.pl. */ + { - "%binary" return PERCENT_NONASSOC; - "%debug" return PERCENT_DEBUG; - "%define" return PERCENT_DEFINE; - "%defines" return PERCENT_DEFINES; - "%destructor" token_type = PERCENT_DESTRUCTOR; BEGIN SC_PRE_CODE; - "%dprec" return PERCENT_DPREC; - "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE; - "%expect" return PERCENT_EXPECT; - "%file-prefix" return PERCENT_FILE_PREFIX; + "%binary" return PERCENT_NONASSOC; + "%code" return PERCENT_CODE; + "%debug" RETURN_PERCENT_FLAG("parse.trace"); + "%default"[-_]"prec" return PERCENT_DEFAULT_PREC; + "%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; + "%expect"[-_]"rr" return PERCENT_EXPECT_RR; + "%file-prefix" return PERCENT_FILE_PREFIX; "%fixed"[-_]"output"[-_]"files" return PERCENT_YACC; - "%glr-parser" return PERCENT_GLR_PARSER; - "%left" return PERCENT_LEFT; - "%lex-param" token_type = PERCENT_LEX_PARAM; BEGIN SC_PRE_CODE; - "%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" token_type = PERCENT_PARSE_PARAM; BEGIN SC_PRE_CODE; - "%prec" rule_length--; return PERCENT_PREC; - "%printer" token_type = PERCENT_PRINTER; BEGIN SC_PRE_CODE; - "%pure"[-_]"parser" return PERCENT_PURE_PARSER; - "%right" return PERCENT_RIGHT; - "%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" token_type = PERCENT_UNION; BEGIN SC_PRE_CODE; - "%verbose" return PERCENT_VERBOSE; - "%yacc" return PERCENT_YACC; + "%initial-action" return PERCENT_INITIAL_ACTION; + "%glr-parser" return PERCENT_GLR_PARSER; + "%language" return PERCENT_LANGUAGE; + "%left" return PERCENT_LEFT; + "%lex-param" RETURN_PERCENT_PARAM(lex); + "%locations" RETURN_PERCENT_FLAG("locations"); + "%merge" return PERCENT_MERGE; + "%name"[-_]"prefix" return PERCENT_NAME_PREFIX; + "%no"[-_]"default"[-_]"prec" return PERCENT_NO_DEFAULT_PREC; + "%no"[-_]"lines" return PERCENT_NO_LINES; + "%nonassoc" return PERCENT_NONASSOC; + "%nondeterministic-parser" return PERCENT_NONDETERMINISTIC_PARSER; + "%nterm" return PERCENT_NTERM; + "%output" return PERCENT_OUTPUT; + "%param" RETURN_PERCENT_PARAM(both); + "%parse-param" RETURN_PERCENT_PARAM(parse); + "%prec" return PERCENT_PREC; + "%precedence" return PERCENT_PRECEDENCE; + "%printer" return PERCENT_PRINTER; + "%pure"[-_]"parser" RETURN_PERCENT_FLAG("api.pure"); + "%require" return PERCENT_REQUIRE; + "%right" return PERCENT_RIGHT; + "%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 (*loc, _("invalid directive: %s"), quote (yytext)); + complain (loc, complaint, _("invalid directive: %s"), quote (yytext)); } "=" return EQUAL; - "|" rule_length = 0; return PIPE; + "|" return PIPE; ";" return SEMICOLON; {id} { - val->symbol = symbol_get (yytext, *loc); + val->uniqstr = uniqstr_new (yytext); id_loc = *loc; - rule_length++; + bracketed_id_str = NULL; BEGIN SC_AFTER_IDENTIFIER; } {int} { - unsigned long num; - set_errno (0); - num = strtoul (yytext, 0, 10); - if (INT_MAX < num || get_errno ()) - { - complain_at (*loc, _("integer out of range: %s"), quote (yytext)); - num = INT_MAX; - } - val->integer = num; + val->integer = scan_integer (yytext, 10, *loc); return INT; } + 0[xX][0-9abcdefABCDEF]+ { + val->integer = scan_integer (yytext, 16, *loc); + return INT; + } + + /* Identifiers may not start with a digit. Yet, don't silently + accept "1FOO" as "1 FOO". */ + {int}{id} { + complain (loc, complaint, _("invalid identifier: %s"), quote (yytext)); + } - /* Characters. We don't check there is only one. */ - "'" STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER; + /* Characters. */ + "'" token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER; /* Strings. */ - "\"" STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_STRING; + "\"" token_start = loc->start; BEGIN SC_ESCAPED_STRING; /* Prologue. */ "%{" code_start = loc->start; BEGIN SC_PROLOGUE; @@ -252,146 +282,313 @@ splice (\\[ \f\t\v]*\n)* /* Code in between braces. */ "{" { STRING_GROW; - token_type = BRACED_CODE; - braces_level = 0; + nesting = 0; code_start = loc->start; BEGIN SC_BRACED_CODE; } + /* Semantic predicate. */ + "%?"[ \f\n\t\v]*"{" { + nesting = 0; + code_start = loc->start; + BEGIN SC_PREDICATE; + } + /* A type. */ + "<*>" return TAG_ANY; + "<>" return TAG_NONE; "<"{tag}">" { obstack_grow (&obstack_for_string, yytext + 1, yyleng - 2); STRING_FINISH; val->uniqstr = uniqstr_new (last_string); STRING_FREE; - return TYPE; + return TAG; + } + "<" { + nesting = 0; + token_start = loc->start; + BEGIN SC_TAG; } "%%" { static int percent_percent_count; if (++percent_percent_count == 2) - { - code_start = loc->start; - BEGIN SC_EPILOGUE; - } + BEGIN SC_EPILOGUE; return PERCENT_PERCENT; } + "[" { + bracketed_id_str = NULL; + bracketed_id_start = loc->start; + bracketed_id_context_state = YY_START; + BEGIN SC_BRACKETED_ID; + } + . { - complain_at (*loc, _("invalid character: %s"), quote (yytext)); + complain (loc, complaint, _("invalid character: %s"), quote (yytext)); + } + + <> { + loc->start = loc->end = scanner_cursor; + yyterminate (); } } + /*--------------------------------------------------------------. + | Supporting \0 complexifies our implementation for no expected | + | added value. | + `--------------------------------------------------------------*/ + + +{ + \0 complain (loc, complaint, _("invalid null character")); +} + + /*-----------------------------------------------------------------. | Scanning after an identifier, checking whether a colon is next. | `-----------------------------------------------------------------*/ { + "[" { + if (bracketed_id_str) + { + ROLLBACK_CURRENT_TOKEN; + BEGIN SC_RETURN_BRACKETED_ID; + *loc = id_loc; + return ID; + } + else + { + bracketed_id_start = loc->start; + bracketed_id_context_state = YY_START; + BEGIN SC_BRACKETED_ID; + } + } ":" { - rule_length = 0; + BEGIN (bracketed_id_str ? SC_RETURN_BRACKETED_ID : INITIAL); *loc = id_loc; - BEGIN INITIAL; return ID_COLON; } . { - scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0); - yyless (0); + ROLLBACK_CURRENT_TOKEN; + BEGIN (bracketed_id_str ? SC_RETURN_BRACKETED_ID : INITIAL); *loc = id_loc; - BEGIN INITIAL; return ID; } <> { + BEGIN (bracketed_id_str ? SC_RETURN_BRACKETED_ID : INITIAL); *loc = id_loc; - BEGIN INITIAL; return ID; } } + /*--------------------------------. + | Scanning bracketed identifiers. | + `--------------------------------*/ + + +{ + {id} { + if (bracketed_id_str) + { + complain (loc, complaint, + _("unexpected identifier in bracketed name: %s"), + quote (yytext)); + } + else + { + bracketed_id_str = uniqstr_new (yytext); + bracketed_id_loc = *loc; + } + } + "]" { + BEGIN bracketed_id_context_state; + if (bracketed_id_str) + { + if (INITIAL == bracketed_id_context_state) + { + val->uniqstr = bracketed_id_str; + bracketed_id_str = 0; + *loc = bracketed_id_loc; + return BRACKETED_ID; + } + } + else + complain (loc, complaint, _("an identifier expected")); + } + . { + complain (loc, complaint, _("invalid character in bracketed name: %s"), + quote (yytext)); + } + <> { + BEGIN bracketed_id_context_state; + unexpected_eof (bracketed_id_start, "]"); + } +} + + +{ + . { + ROLLBACK_CURRENT_TOKEN; + val->uniqstr = bracketed_id_str; + bracketed_id_str = 0; + *loc = bracketed_id_loc; + BEGIN INITIAL; + return BRACKETED_ID; + } +} + /*---------------------------------------------------------------. - | Scanning a Yacc comment. The initial `/ *' is already eaten. | + | Scanning a Yacc comment. The initial '/ *' is already eaten. | `---------------------------------------------------------------*/ { "*/" BEGIN context_state; - .|\n ; - <> unexpected_end_of_file (token_start, "*/"); + .|\n ; + <> unexpected_eof (token_start, "*/"); BEGIN context_state; } /*------------------------------------------------------------. - | Scanning a C comment. The initial `/ *' is already eaten. | + | Scanning a C comment. The initial '/ *' is already eaten. | `------------------------------------------------------------*/ { "*"{splice}"/" STRING_GROW; BEGIN context_state; - <> unexpected_end_of_file (token_start, "*/"); + <> unexpected_eof (token_start, "*/"); BEGIN context_state; } /*--------------------------------------------------------------. - | Scanning a line comment. The initial `//' is already eaten. | + | Scanning a line comment. The initial '//' is already eaten. | `--------------------------------------------------------------*/ { - "\n" STRING_GROW; BEGIN context_state; - {splice} STRING_GROW; - <> BEGIN context_state; + "\n" STRING_GROW; BEGIN context_state; + {splice} STRING_GROW; + <> BEGIN context_state; } - /*----------------------------------------------------------------. - | Scanning a C string, including its escapes. The initial `"' is | - | already eaten. | - `----------------------------------------------------------------*/ + /*------------------------------------------------. + | Scanning a Bison string, including its escapes. | + | The initial quote is already eaten. | + `------------------------------------------------*/ { - "\"" { - STRING_GROW; + "\""|"\n" { + if (yytext[0] == '\n') + unexpected_newline (token_start, "\""); + STRING_FINISH; + loc->start = token_start; + val->chars = last_string; + BEGIN INITIAL; + return STRING; + } + <> { + unexpected_eof (token_start, "\""); STRING_FINISH; loc->start = token_start; val->chars = last_string; - rule_length++; BEGIN INITIAL; return STRING; } - - .|\n STRING_GROW; - <> unexpected_end_of_file (token_start, "\""); } - /*---------------------------------------------------------------. - | Scanning a C character, decoding its escapes. The initial "'" | - | is already eaten. | - `---------------------------------------------------------------*/ + /*----------------------------------------------------------. + | Scanning a Bison character literal, decoding its escapes. | + | The initial quote is already eaten. | + `----------------------------------------------------------*/ { - "'" { - unsigned char last_string_1; - STRING_GROW; + "'"|"\n" { STRING_FINISH; loc->start = token_start; - val->symbol = symbol_get (last_string, *loc); - symbol_class_set (val->symbol, token_sym, *loc); - last_string_1 = last_string[1]; - symbol_user_token_number_set (val->symbol, last_string_1, *loc); + val->character = last_string[0]; + { + /* FIXME: Eventually, make these errors. */ + if (last_string[0] == '\0') + { + complain (loc, Wother, _("empty character literal")); + /* '\0' seems dangerous even if we are about to complain. */ + val->character = '\''; + } + else if (last_string[1] != '\0') + complain (loc, Wother, + _("extra characters in character literal")); + } + if (yytext[0] == '\n') + unexpected_newline (token_start, "'"); STRING_FREE; - rule_length++; BEGIN INITIAL; - return ID; + return CHAR; + } + <> { + STRING_FINISH; + loc->start = token_start; + val->character = last_string[0]; + { + /* FIXME: Eventually, make these errors. */ + if (last_string[0] == '\0') + { + complain (loc, Wother, _("empty character literal")); + /* '\0' seems dangerous even if we are about to complain. */ + val->character = '\''; + } + else if (last_string[1] != '\0') + complain (loc, Wother, + _("extra characters in character literal")); + } + unexpected_eof (token_start, "'"); + STRING_FREE; + BEGIN INITIAL; + return CHAR; } - - .|\n STRING_GROW; - <> unexpected_end_of_file (token_start, "'"); } + /*-----------------------------------------------------------. + | Scanning a Bison nested tag. The initial angle bracket is | + | already eaten. | + `-----------------------------------------------------------*/ + + +{ + ">" { + --nesting; + if (nesting < 0) + { + STRING_FINISH; + loc->start = token_start; + val->uniqstr = uniqstr_new (last_string); + STRING_FREE; + BEGIN INITIAL; + return TAG; + } + STRING_GROW; + } + + [^<>]+ STRING_GROW; + "<"+ STRING_GROW; nesting += yyleng; + + <> { + unexpected_eof (token_start, ">"); + STRING_FINISH; + loc->start = token_start; + val->uniqstr = uniqstr_new (last_string); + STRING_FREE; + BEGIN INITIAL; + return TAG; + } +} /*----------------------------. | Decode escaped characters. | @@ -400,71 +597,76 @@ splice (\\[ \f\t\v]*\n)* { \\[0-7]{1,3} { - unsigned long c = strtoul (yytext + 1, 0, 8); - if (UCHAR_MAX < c) - complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext)); + unsigned long int c = strtoul (yytext + 1, NULL, 8); + if (!c || UCHAR_MAX < c) + complain (loc, complaint, _("invalid number after \\-escape: %s"), + yytext+1); else obstack_1grow (&obstack_for_string, c); } \\x[0-9abcdefABCDEF]+ { - unsigned long c; - set_errno (0); - c = strtoul (yytext + 2, 0, 16); - if (UCHAR_MAX < c || get_errno ()) - complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext)); + verify (UCHAR_MAX < ULONG_MAX); + unsigned long int c = strtoul (yytext + 2, NULL, 16); + if (!c || UCHAR_MAX < c) + complain (loc, complaint, _("invalid number after \\-escape: %s"), + yytext+1); else obstack_1grow (&obstack_for_string, c); } - \\a obstack_1grow (&obstack_for_string, '\a'); - \\b obstack_1grow (&obstack_for_string, '\b'); - \\f obstack_1grow (&obstack_for_string, '\f'); - \\n obstack_1grow (&obstack_for_string, '\n'); - \\r obstack_1grow (&obstack_for_string, '\r'); - \\t obstack_1grow (&obstack_for_string, '\t'); - \\v obstack_1grow (&obstack_for_string, '\v'); + \\a obstack_1grow (&obstack_for_string, '\a'); + \\b obstack_1grow (&obstack_for_string, '\b'); + \\f obstack_1grow (&obstack_for_string, '\f'); + \\n obstack_1grow (&obstack_for_string, '\n'); + \\r obstack_1grow (&obstack_for_string, '\r'); + \\t obstack_1grow (&obstack_for_string, '\t'); + \\v obstack_1grow (&obstack_for_string, '\v'); /* \\[\"\'?\\] would be shorter, but it confuses xgettext. */ \\("\""|"'"|"?"|"\\") obstack_1grow (&obstack_for_string, yytext[1]); \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} { int c = convert_ucn_to_byte (yytext); - if (c < 0) - complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext)); + if (c <= 0) + complain (loc, complaint, _("invalid number after \\-escape: %s"), + yytext+1); else obstack_1grow (&obstack_for_string, c); } - \\(.|\n) { - complain_at (*loc, _("unrecognized escape sequence: %s"), quote (yytext)); - STRING_GROW; + \\(.|\n) { + char const *p = yytext + 1; + /* Quote only if escaping won't make the character visible. */ + if (c_isspace ((unsigned char) *p) && c_isprint ((unsigned char) *p)) + p = quote (p); + else + p = quotearg_style_mem (escape_quoting_style, p, 1); + complain (loc, complaint, _("invalid character after \\-escape: %s"), + p); } } + /*--------------------------------------------. + | Scanning user-code characters and strings. | + `--------------------------------------------*/ - /*----------------------------------------------------------. - | Scanning a C character without decoding its escapes. The | - | initial "'" is already eaten. | - `----------------------------------------------------------*/ + +{ + {splice}|\\{splice}[^\n\[\]] STRING_GROW; +} { - "'" STRING_GROW; BEGIN context_state; - \\{splice}[^$@\[\]] STRING_GROW; - <> unexpected_end_of_file (token_start, "'"); + "'" STRING_GROW; BEGIN context_state; + \n unexpected_newline (token_start, "'"); BEGIN context_state; + <> unexpected_eof (token_start, "'"); BEGIN context_state; } - - /*----------------------------------------------------------------. - | Scanning a C string, without decoding its escapes. The initial | - | `"' is already eaten. | - `----------------------------------------------------------------*/ - { - "\"" STRING_GROW; BEGIN context_state; - \\{splice}[^$@\[\]] STRING_GROW; - <> unexpected_end_of_file (token_start, "\""); + "\"" STRING_GROW; BEGIN context_state; + \n unexpected_newline (token_start, "\""); BEGIN context_state; + <> unexpected_eof (token_start, "\""); BEGIN context_state; } @@ -472,7 +674,7 @@ splice (\\[ \f\t\v]*\n)* | Strings, comments etc. can be found in user code. | `---------------------------------------------------*/ - + { "'" { STRING_GROW; @@ -500,91 +702,65 @@ splice (\\[ \f\t\v]*\n)* } - /*---------------------------------------------------------------. - | Scanning after %union etc., possibly followed by white space. | - | For %union only, allow arbitrary C code to appear before the | - | following brace, as an extension to POSIX. | - `---------------------------------------------------------------*/ - + /*-----------------------------------------------------------. + | Scanning some code in braces (actions, predicates). The | + | initial "{" is already eaten. | + `-----------------------------------------------------------*/ + + { - . { - bool valid = yytext[0] == '{' || token_type == PERCENT_UNION; - scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0); - yyless (0); + "{"|"<"{splice}"%" STRING_GROW; nesting++; + "%"{splice}">" STRING_GROW; nesting--; - if (valid) - { - braces_level = -1; - code_start = loc->start; - BEGIN SC_BRACED_CODE; - } - else - { - complain_at (*loc, _("missing `{' in `%s'"), - token_name (token_type)); - obstack_sgrow (&obstack_for_string, "{}"); - STRING_FINISH; - val->chars = last_string; - BEGIN INITIAL; - return token_type; - } + /* Tokenize '<<%' correctly (as '<<' '%') rather than incorrrectly + (as '<' '<%'). */ + "<"{splice}"<" STRING_GROW; + + <> { + int token = (YY_START == SC_BRACED_CODE) ? BRACED_CODE : BRACED_PREDICATE; + unexpected_eof (code_start, "}"); + STRING_FINISH; + loc->start = code_start; + val->code = last_string; + BEGIN INITIAL; + return token; } } - - /*---------------------------------------------------------------. - | Scanning some code in braces (%union and actions). The initial | - | "{" is already eaten. | - `---------------------------------------------------------------*/ - { - "{"|"<"{splice}"%" STRING_GROW; braces_level++; - "%"{splice}">" STRING_GROW; braces_level--; "}" { - bool outer_brace = --braces_level < 0; - - /* As an undocumented Bison extension, append `;' before the last - brace in braced code, so that the user code can omit trailing - `;'. But do not append `;' if emulating Yacc, since Yacc does - not append one. - - FIXME: Bison should warn if a semicolon seems to be necessary - here, and should omit the semicolon if it seems unnecessary - (e.g., after ';', '{', or '}', each followed by comments or - white space). Such a warning shouldn't depend on --yacc; it - should depend on a new --pedantic option, which would cause - Bison to warn if it detects an extension to POSIX. --pedantic - should also diagnose other Bison extensions like %yacc. - Perhaps there should also be a GCC-style --pedantic-errors - option, so that such warnings are diagnosed as errors. */ - if (outer_brace && ! yacc_flag) - obstack_1grow (&obstack_for_string, ';'); - obstack_1grow (&obstack_for_string, '}'); - if (outer_brace) + --nesting; + if (nesting < 0) { - STRING_FINISH; - rule_length++; - loc->start = code_start; - val->chars = last_string; - BEGIN INITIAL; - return token_type; + STRING_FINISH; + loc->start = code_start; + val->code = last_string; + BEGIN INITIAL; + return BRACED_CODE; } } - - /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly - (as `<' `<%'). */ - "<"{splice}"<" STRING_GROW; - - "$"("<"{tag}">")?(-?[0-9]+|"$") handle_dollar (token_type, yytext, *loc); - "@"(-?[0-9]+|"$") handle_at (token_type, yytext, *loc); - - <> unexpected_end_of_file (code_start, "}"); } + +{ + "}" { + --nesting; + if (nesting < 0) + { + STRING_FINISH; + loc->start = code_start; + val->code = last_string; + BEGIN INITIAL; + return BRACED_PREDICATE; + } + else + obstack_1grow (&obstack_for_string, '}'); + } +} /*--------------------------------------------------------------. | Scanning some prologue: from "%{" (already scanned) to "%}". | @@ -600,7 +776,14 @@ splice (\\[ \f\t\v]*\n)* return PROLOGUE; } - <> unexpected_end_of_file (code_start, "%}"); + <> { + unexpected_eof (code_start, "%}"); + STRING_FINISH; + loc->start = code_start; + val->chars = last_string; + BEGIN INITIAL; + return PROLOGUE; + } } @@ -621,60 +804,15 @@ splice (\\[ \f\t\v]*\n)* } - /*----------------------------------------------------------------. - | By default, grow the string obstack with the input, escaping M4 | - | quoting characters. | - `----------------------------------------------------------------*/ - - -{ - \$ obstack_sgrow (&obstack_for_string, "$]["); - \@ obstack_sgrow (&obstack_for_string, "@@"); - \[ obstack_sgrow (&obstack_for_string, "@{"); - \] obstack_sgrow (&obstack_for_string, "@}"); - .|\n STRING_GROW; -} + /*-----------------------------------------------------. + | By default, grow the string obstack with the input. | + `-----------------------------------------------------*/ +. | + \n STRING_GROW; %% -/* Set *LOC and adjust scanner cursor to account for token TOKEN of - size SIZE. */ - -static void -adjust_location (location *loc, char const *token, size_t size) -{ - int line = scanner_cursor.line; - int column = scanner_cursor.column; - char const *p0 = token; - char const *p = token; - char const *lim = token + size; - - loc->start = scanner_cursor; - - for (p = token; p < lim; p++) - switch (*p) - { - 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; - } - - scanner_cursor.line = line; - scanner_cursor.column = column + mbsnwidth (p0, p - p0, 0); - - loc->end = scanner_cursor; -} - - /* 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. */ @@ -682,222 +820,79 @@ adjust_location (location *loc, char const *token, size_t size) static size_t no_cr_read (FILE *fp, char *buf, size_t size) { - size_t s = fread (buf, 1, size, fp); - if (s) + size_t bytes_read = fread (buf, 1, size, fp); + if (bytes_read) { - char *w = memchr (buf, '\r', s); + char *w = memchr (buf, '\r', bytes_read); 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; -} - - -/*------------------------------------------------------------------. -| TEXT is pointing to a wannabee semantic value (i.e., a `$'). | -| | -| Possible inputs: $[]($|integer) | -| | -| Output to OBSTACK_FOR_STRING a reference to this semantic value. | -`------------------------------------------------------------------*/ - -static inline bool -handle_action_dollar (char *text, location loc) -{ - const char *type_name = NULL; - char *cp = text + 1; - - if (! current_rule) - return false; - - /* 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, loc, 0); - if (!type_name && typed) - complain_at (loc, _("$$ of `%s' has no declared type"), - current_rule->sym->tag); - if (!type_name) - type_name = ""; - obstack_fgrow1 (&obstack_for_string, - "]b4_lhs_value([%s])[", type_name); - } - else - { - long num; - set_errno (0); - num = strtol (cp, 0, 10); - - if (INT_MIN <= num && num <= rule_length && ! get_errno ()) - { - int n = num; - if (!type_name && n > 0) - type_name = symbol_list_n_type_name_get (current_rule, loc, n); - if (!type_name && typed) - complain_at (loc, _("$%d of `%s' has no declared type"), - n, current_rule->sym->tag); - if (!type_name) - type_name = ""; - obstack_fgrow3 (&obstack_for_string, - "]b4_rhs_value([%d], [%d], [%s])[", - rule_length, n, type_name); - } - else - complain_at (loc, _("integer out of range: %s"), quote (text)); + { + char const *r = ++w; + char const *lim = buf + bytes_read; + + for (;;) + { + /* Found an '\r'. Treat it like '\n', but ignore any + '\n' that immediately follows. */ + w[-1] = '\n'; + if (r == lim) + { + int ch = getc (fp); + if (ch != '\n' && ungetc (ch, fp) != ch) + break; + } + else if (*r == '\n') + r++; + + /* Copy until the next '\r'. */ + do + { + if (r == lim) + return w - buf; + } + while ((*w++ = *r++) != '\r'); + } + + return w - buf; + } } - return true; + return bytes_read; } -/*-----------------------------------------------------------------. -| Dispatch onto handle_action_dollar, or handle_destructor_dollar, | -| depending upon TOKEN_TYPE. | -`-----------------------------------------------------------------*/ - -static void -handle_dollar (int token_type, char *text, location loc) -{ - switch (token_type) - { - case BRACED_CODE: - if (handle_action_dollar (text, loc)) - return; - break; - - case PERCENT_DESTRUCTOR: - case PERCENT_PRINTER: - if (text[1] == '$') - { - obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar["); - return; - } - break; - - default: - break; - } - - complain_at (loc, _("invalid value: %s"), quote (text)); -} - /*------------------------------------------------------. -| TEXT is a location token (i.e., a `@...'). Output to | -| OBSTACK_FOR_STRING a reference to this location. | +| Scan NUMBER for a base-BASE integer at location LOC. | `------------------------------------------------------*/ -static inline bool -handle_action_at (char *text, location loc) +static unsigned long int +scan_integer (char const *number, int base, location loc) { - char *cp = text + 1; - locations_flag = 1; + verify (INT_MAX < ULONG_MAX); + unsigned long int num = strtoul (number, NULL, base); - if (! current_rule) - return false; - - if (*cp == '$') - obstack_sgrow (&obstack_for_string, "]b4_lhs_location["); - else + if (INT_MAX < num) { - long num; - set_errno (0); - num = strtol (cp, 0, 10); - - if (INT_MIN <= num && num <= rule_length && ! get_errno ()) - { - int n = num; - obstack_fgrow2 (&obstack_for_string, "]b4_rhs_location([%d], [%d])[", - rule_length, n); - } - else - complain_at (loc, _("integer out of range: %s"), quote (text)); + complain (&loc, complaint, _("integer out of range: %s"), + quote (number)); + num = INT_MAX; } - return true; -} - - -/*-------------------------------------------------------------------. -| Dispatch onto handle_action_at, or handle_destructor_at, depending | -| upon CODE_KIND. | -`-------------------------------------------------------------------*/ - -static void -handle_at (int token_type, char *text, location loc) -{ - switch (token_type) - { - case BRACED_CODE: - handle_action_at (text, loc); - return; - - case PERCENT_DESTRUCTOR: - case PERCENT_PRINTER: - if (text[1] == '$') - { - obstack_sgrow (&obstack_for_string, "]b4_at_dollar["); - return; - } - break; - - default: - break; - } - - complain_at (loc, _("invalid value: %s"), quote (text)); + return num; } /*------------------------------------------------------------------. | 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. | +| to a single-byte character. | `------------------------------------------------------------------*/ static int convert_ucn_to_byte (char const *ucn) { - unsigned long code = strtoul (ucn + 2, 0, 16); + verify (UCHAR_MAX <= INT_MAX); + unsigned long int code = strtoul (ucn + 2, NULL, 16); /* FIXME: Currently we assume Unicode-compatible unibyte characters on ASCII hosts (i.e., Latin-1 on hosts with 8-bit bytes). On @@ -918,22 +913,22 @@ convert_ucn_to_byte (char const *ucn) 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, - ' ', '!', '"', '#', '$', '%', '&', '\'', - '(', ')', '*', '+', ',', '-', '.', '/', - '0', '1', '2', '3', '4', '5', '6', '7', - '8', '9', ':', ';', '<', '=', '>', '?', - '@', 'A', 'B', 'C', 'D', 'E', 'F', 'G', - 'H', 'I', 'J', 'K', 'L', 'M', 'N', 'O', - 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', - 'X', 'Y', 'Z', '[', '\\', ']', '^', '_', - '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', - 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', - 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', - 'x', 'y', 'z', '{', '|', '}', '~' + '\0', -1, -1, -1, -1, -1, -1, '\a', + '\b', '\t', '\n', '\v', '\f', '\r', -1, -1, + -1, -1, -1, -1, -1, -1, -1, -1, + -1, -1, -1, -1, -1, -1, -1, -1, + ' ', '!', '"', '#', '$', '%', '&', '\'', + '(', ')', '*', '+', ',', '-', '.', '/', + '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; @@ -945,44 +940,66 @@ convert_ucn_to_byte (char const *ucn) /*----------------------------------------------------------------. -| Handle `#line INT "FILE"'. ARGS has already skipped `#line '. | +| Handle '#line INT "FILE"'. ARGS has already skipped '#line '. | `----------------------------------------------------------------*/ static void -handle_syncline (char *args) +handle_syncline (char *args, location loc) { - int lineno = strtol (args, &args, 10); - const char *file = NULL; - file = strchr (args, '"') + 1; - *strchr (file, '"') = 0; - scanner_cursor.file = current_file = xstrdup (file); - scanner_cursor.line = lineno; - scanner_cursor.column = 1; + char *after_num; + unsigned long int lineno = strtoul (args, &after_num, 10); + char *file = strchr (after_num, '"') + 1; + *strchr (file, '"') = '\0'; + if (INT_MAX <= lineno) + { + complain (&loc, Wother, _("line number overflow")); + lineno = INT_MAX; + } + current_file = uniqstr_new (file); + boundary_set (&scanner_cursor, current_file, lineno, 1); +} + + +/*----------------------------------------------------------------. +| For a token or comment starting at START, report message MSGID, | +| which should say that an end marker was found before | +| the expected TOKEN_END. | +`----------------------------------------------------------------*/ + +static void +unexpected_end (boundary start, char const *msgid, char const *token_end) +{ + location loc; + loc.start = start; + loc.end = scanner_cursor; + token_end = quote (token_end); + // Instead of '\'', display "'". + if (STREQ (token_end, "'\\''")) + token_end = "\"'\""; + complain (&loc, complaint, _(msgid), token_end); } /*------------------------------------------------------------------------. | Report an unexpected EOF in a token or comment starting at START. | | An end of file was encountered and the expected TOKEN_END was missing. | -| After reporting the problem, pretend that TOKEN_END was found. | `------------------------------------------------------------------------*/ static void -unexpected_end_of_file (boundary start, char const *token_end) +unexpected_eof (boundary start, char const *token_end) { - size_t i = strlen (token_end); + unexpected_end (start, N_("missing %s at end of file"), token_end); +} - location loc; - loc.start = start; - loc.end = scanner_cursor; - complain_at (loc, _("missing `%s' at end of file"), token_end); - /* Adjust scanner cursor so that any later message does not count - the characters about to be inserted. */ - scanner_cursor.column -= i; +/*----------------------------------------. +| Likewise, but for unexpected newlines. | +`----------------------------------------*/ - while (i != 0) - unput (token_end[--i]); +static void +unexpected_newline (boundary start, char const *token_end) +{ + unexpected_end (start, N_("missing %s at end of line"), token_end); } @@ -991,7 +1008,7 @@ unexpected_end_of_file (boundary start, char const *token_end) `-------------------------*/ void -scanner_initialize (void) +gram_scanner_initialize (void) { obstack_init (&obstack_for_string); } @@ -1002,9 +1019,9 @@ scanner_initialize (void) `-----------------------------------------------*/ void -scanner_free (void) +gram_scanner_free (void) { obstack_free (&obstack_for_string, 0); /* Reclaim Flex's buffers. */ - yy_delete_buffer (YY_CURRENT_BUFFER); + yylex_destroy (); }