X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/60457f308d3cbe377472b610aafcb86e194081a4..50977317cd10df482c3b35f61e9ee4522eee3c3b:/src/parse-gram.y diff --git a/src/parse-gram.y b/src/parse-gram.y index 8d745491..6e58835a 100644 --- a/src/parse-gram.y +++ b/src/parse-gram.y @@ -20,6 +20,7 @@ #include #include "system.h" +#include "c-ctype.h" #include "complain.h" #include "conflicts.h" #include "files.h" @@ -30,14 +31,16 @@ #include "quotearg.h" #include "reader.h" #include "symlist.h" +#include "symtab.h" #include "scan-gram.h" #include "scan-code.h" +#include "xmemdup0.h" #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N) static YYLTYPE lloc_default (YYLTYPE const *, int); #define YY_LOCATION_PRINT(File, Loc) \ - location_print (File, Loc) + location_print (File, Loc) static void version_check (location const *loc, char const *version); @@ -45,61 +48,54 @@ static void version_check (location const *loc, char const *version); FIXME: depends on the undocumented availability of YYLLOC. */ #undef yyerror #define yyerror(Msg) \ - gram_error (&yylloc, Msg) + gram_error (&yylloc, Msg) static void gram_error (location const *, char const *); static char const *char_name (char); +%} -/** Add a lex-param or a parse-param. - * - * \param type \a lex_param or \a parse_param - * \param decl the formal argument - * \param loc the location in the source. - */ -static void add_param (char const *type, char *decl, location loc); - - -static symbol_class current_class = unknown_sym; -static uniqstr current_type = NULL; -static symbol *current_lhs_symbol; -static location current_lhs_location; -static named_ref *current_lhs_named_ref; -static int current_prec = 0; - -/** Set the new current left-hand side symbol, possibly common - * to several right-hand side parts of rule. - */ -static -void -current_lhs(symbol *sym, location loc, named_ref *ref) +%code { - current_lhs_symbol = sym; - current_lhs_location = loc; - /* In order to simplify memory management, named references for lhs - are always assigned by deep copy into the current symbol_list - node. This is because a single named-ref in the grammar may - result in several uses when the user factors lhs between several - rules using "|". Therefore free the parser's original copy. */ - free (current_lhs_named_ref); - current_lhs_named_ref = ref; -} - + static int current_prec = 0; + static location current_lhs_location; + static named_ref *current_lhs_named_ref; + static symbol *current_lhs_symbol; + static symbol_class current_class = unknown_sym; + static uniqstr current_type = NULL; + + /** Set the new current left-hand side symbol, possibly common + * to several right-hand side parts of rule. + */ + static + void + current_lhs(symbol *sym, location loc, named_ref *ref) + { + current_lhs_symbol = sym; + current_lhs_location = loc; + /* In order to simplify memory management, named references for lhs + are always assigned by deep copy into the current symbol_list + node. This is because a single named-ref in the grammar may + result in several uses when the user factors lhs between several + rules using "|". Therefore free the parser's original copy. */ + free (current_lhs_named_ref); + current_lhs_named_ref = ref; + } -#define YYTYPE_INT16 int_fast16_t -#define YYTYPE_INT8 int_fast8_t -#define YYTYPE_UINT16 uint_fast16_t -#define YYTYPE_UINT8 uint_fast8_t -%} + #define YYTYPE_INT16 int_fast16_t + #define YYTYPE_INT8 int_fast8_t + #define YYTYPE_UINT16 uint_fast16_t + #define YYTYPE_UINT8 uint_fast8_t +} -%debug -%verbose -%defines -%locations -%pure-parser -%error-verbose +%define api.prefix "gram_" +%define api.pure +%define locations +%define parse.error verbose %define parse.lac full -%name-prefix="gram_" +%define parse.trace +%defines %expect 0 +%verbose %initial-action { @@ -111,15 +107,15 @@ current_lhs(symbol *sym, location loc, named_ref *ref) %union { + assoc assoc; + char *code; + char const *chars; + int integer; + named_ref *named_ref; symbol *symbol; symbol_list *list; - int integer; - char const *chars; - char *code; - assoc assoc; uniqstr uniqstr; unsigned char character; - named_ref *named_ref; }; /* Define the tokens together with their human representation. */ @@ -137,40 +133,36 @@ current_lhs(symbol *sym, location loc, named_ref *ref) %token PERCENT_LEFT "%left" %token PERCENT_RIGHT "%right" %token PERCENT_NONASSOC "%nonassoc" +%token PERCENT_PRECEDENCE "%precedence" %token PERCENT_PREC "%prec" %token PERCENT_DPREC "%dprec" %token PERCENT_MERGE "%merge" - /*----------------------. | Global Declarations. | `----------------------*/ %token PERCENT_CODE "%code" - PERCENT_DEBUG "%debug" PERCENT_DEFAULT_PREC "%default-prec" PERCENT_DEFINE "%define" PERCENT_DEFINES "%defines" PERCENT_ERROR_VERBOSE "%error-verbose" PERCENT_EXPECT "%expect" - PERCENT_EXPECT_RR "%expect-rr" + PERCENT_EXPECT_RR "%expect-rr" + PERCENT_FLAG "%" PERCENT_FILE_PREFIX "%file-prefix" PERCENT_GLR_PARSER "%glr-parser" PERCENT_INITIAL_ACTION "%initial-action" PERCENT_LANGUAGE "%language" - PERCENT_LEX_PARAM "%lex-param" - PERCENT_LOCATIONS "%locations" PERCENT_NAME_PREFIX "%name-prefix" PERCENT_NO_DEFAULT_PREC "%no-default-prec" PERCENT_NO_LINES "%no-lines" PERCENT_NONDETERMINISTIC_PARSER - "%nondeterministic-parser" + "%nondeterministic-parser" PERCENT_OUTPUT "%output" - PERCENT_PARSE_PARAM "%parse-param" - PERCENT_PURE_PARSER "%pure-parser" - PERCENT_REQUIRE "%require" + PERCENT_REQUIRE "%require" PERCENT_SKELETON "%skeleton" PERCENT_START "%start" PERCENT_TOKEN_TABLE "%token-table" @@ -179,6 +171,7 @@ current_lhs(symbol *sym, location loc, named_ref *ref) ; %token BRACED_CODE "{...}" +%token BRACED_PREDICATE "%?{...}" %token BRACKETED_ID "[identifier]" %token CHAR "char" %token EPILOGUE "epilogue" @@ -189,39 +182,90 @@ current_lhs(symbol *sym, location loc, named_ref *ref) %token PIPE "|" %token PROLOGUE "%{...%}" %token SEMICOLON ";" -%token TYPE "type" -%token TYPE_TAG_ANY "<*>" -%token TYPE_TAG_NONE "<>" +%token TAG "" +%token TAG_ANY "<*>" +%token TAG_NONE "<>" %type CHAR -%printer { fputs (char_name ($$), stderr); } CHAR +%printer { fputs (char_name ($$), yyo); } CHAR /* braceless is not to be used for rule or symbol actions, as it calls code_props_plain_init. */ %type STRING "%{...%}" EPILOGUE braceless content.opt -%type "{...}" -%printer { fputs (quotearg_style (c_quoting_style, $$), stderr); } - STRING -%printer { fprintf (stderr, "{\n%s\n}", $$); } - braceless content.opt "{...}" "%{...%}" EPILOGUE - -%type BRACKETED_ID ID ID_COLON TYPE variable -%printer { fputs ($$, stderr); } -%printer { fprintf (stderr, "[%s]", $$); } BRACKETED_ID -%printer { fprintf (stderr, "%s:", $$); } ID_COLON -%printer { fprintf (stderr, "<%s>", $$); } TYPE +%type "{...}" "%?{...}" +%printer { fputs (quotearg_style (c_quoting_style, $$), yyo); } + STRING +%printer { fprintf (yyo, "{\n%s\n}", $$); } + braceless content.opt "{...}" "%{...%}" EPILOGUE + +%type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable +%printer { fputs ($$, yyo); } +%printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID +%printer { fprintf (yyo, "%s:", $$); } ID_COLON +%printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG +%printer { fprintf (yyo, "<%s>", $$); } TAG tag %type INT -%printer { fprintf (stderr, "%d", $$); } +%printer { fprintf (yyo, "%d", $$); } %type id id_colon string_as_id symbol symbol.prec -%printer { fprintf (stderr, "%s", $$->tag); } -%printer { fprintf (stderr, "%s:", $$->tag); } id_colon +%printer { fprintf (yyo, "%s", $$->tag); } +%printer { fprintf (yyo, "%s:", $$->tag); } id_colon %type precedence_declarator %type symbols.1 symbols.prec generic_symlist generic_symlist_item %type named_ref.opt +/*---------. +| %param. | +`---------*/ +%code requires +{ +# ifndef PARAM_TYPE +# define PARAM_TYPE + typedef enum + { + param_none = 0, + param_lex = 1 << 0, + param_parse = 1 << 1, + param_both = param_lex | param_parse + } param_type; +# endif +}; +%code +{ + /** Add a lex-param and/or a parse-param. + * + * \param type where to push this formal argument. + * \param decl the formal argument. Destroyed. + * \param loc the location in the source. + */ + static void add_param (param_type type, char *decl, location loc); + static param_type current_param = param_none; +}; +%union +{ + param_type param; +} +%token PERCENT_PARAM "%param"; +%printer +{ + switch ($$) + { +#define CASE(In, Out) \ + case param_ ## In: fputs ("%" #Out, stderr); break + CASE(lex, lex-param); + CASE(parse, parse-param); + CASE(both, param); +#undef CASE + case param_none: aver (false); break; + } +} ; + + + /*==========\ + | Grammar. | + \==========*/ %% input: @@ -229,9 +273,9 @@ input: ; - /*------------------------------------. - | Declarations: before the first %%. | - `------------------------------------*/ + /*------------------------------------. + | Declarations: before the first %%. | + `------------------------------------*/ prologue_declarations: /* Nothing */ @@ -250,7 +294,10 @@ prologue_declaration: plain_code.code, @1); code_scanner_last_string_free (); } -| "%debug" { debug_flag = true; } +| "%" + { + muscle_percent_define_ensure ($1, @1, true); + } | "%define" variable content.opt { muscle_percent_define_insert ($2, @2, $3, @@ -262,11 +309,14 @@ prologue_declaration: defines_flag = true; spec_defines_file = xstrdup ($2); } -| "%error-verbose" { error_verbose = true; } +| "%error-verbose" + { + muscle_percent_define_insert ("parse.error", @1, "verbose", + MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); + } | "%expect" INT { expected_sr_conflicts = $2; } -| "%expect-rr" INT { expected_rr_conflicts = $2; } +| "%expect-rr" INT { expected_rr_conflicts = $2; } | "%file-prefix" STRING { spec_file_prefix = $2; } -| "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */ | "%glr-parser" { nondeterministic_parser = true; @@ -281,36 +331,17 @@ prologue_declaration: muscle_code_grow ("initial_action", action.code, @2); code_scanner_last_string_free (); } -| "%language" STRING { language_argmatch ($2, grammar_prio, @1); } -| "%lex-param" "{...}" { add_param ("lex_param", $2, @2); } -| "%locations" { locations_flag = true; } +| "%language" STRING { language_argmatch ($2, grammar_prio, @1); } | "%name-prefix" STRING { spec_name_prefix = $2; } -| "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */ | "%no-lines" { no_lines_flag = true; } -| "%nondeterministic-parser" { nondeterministic_parser = true; } +| "%nondeterministic-parser" { nondeterministic_parser = true; } | "%output" STRING { spec_outfile = $2; } -| "%output" "=" STRING { spec_outfile = $3; } /* deprecated */ -| "%parse-param" "{...}" { add_param ("parse_param", $2, @2); } -| "%pure-parser" - { - /* %pure-parser is deprecated in favor of `%define api.pure', so use - `%define api.pure' in a backward-compatible manner here. First, don't - complain if %pure-parser is specified multiple times. */ - if (!muscle_find_const ("percent_define(api.pure)")) - muscle_percent_define_insert ("api.pure", @1, "", - MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); - /* In all cases, use api.pure now so that the backend doesn't complain if - the skeleton ignores api.pure, but do warn now if there's a previous - conflicting definition from an actual %define. */ - if (!muscle_percent_define_flag_if ("api.pure")) - muscle_percent_define_insert ("api.pure", @1, "", - MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); - } +| "%param" { current_param = $1; } params { current_param = param_none; } | "%require" STRING { version_check (&@2, $2); } | "%skeleton" STRING { char const *skeleton_user = $2; - if (mbschr (skeleton_user, '/')) + if (strchr (skeleton_user, '/')) { size_t dir_length = strlen (current_file); char *skeleton_build; @@ -337,6 +368,16 @@ prologue_declaration: | /*FIXME: Err? What is this horror doing here? */ ";" ; +params: + params "{...}" { add_param (current_param, $2, @2); } +| "{...}" { add_param (current_param, $1, @1); } +; + + +/*----------------------. +| grammar_declaration. | +`----------------------*/ + grammar_declaration: precedence_declaration | symbol_declaration @@ -344,19 +385,17 @@ grammar_declaration: { grammar_start_symbol_set ($2, @2); } -| "%destructor" "{...}" generic_symlist - { - symbol_list *list; - for (list = $3; list; list = list->next) - symbol_list_destructor_set (list, $2, @2); - symbol_list_free ($3); - } -| "%printer" "{...}" generic_symlist +| code_props_type "{...}" generic_symlist { - symbol_list *list; - for (list = $3; list; list = list->next) - symbol_list_printer_set (list, $2, @2); - symbol_list_free ($3); + code_props code; + code_props_symbol_action_init (&code, $2, @2); + code_props_translate_code (&code); + { + symbol_list *list; + for (list = $3; list; list = list->next) + symbol_list_code_props_set (list, $1, &code); + symbol_list_free ($3); + } } | "%default-prec" { @@ -380,10 +419,17 @@ grammar_declaration: } ; +%type code_props_type; +%union {code_props_type code_type;}; +%printer { fprintf (yyo, "%s", code_props_type_string ($$)); } ; +code_props_type: + "%destructor" { $$ = destructor; } +| "%printer" { $$ = printer; } +; -/*----------* - | %union. | - *----------*/ +/*---------. +| %union. | +`---------*/ %token PERCENT_UNION "%union"; @@ -415,40 +461,41 @@ symbol_declaration: current_class = unknown_sym; current_type = NULL; } -| "%type" TYPE symbols.1 +| "%type" TAG symbols.1 { symbol_list *list; tag_seen = true; for (list = $3; list; list = list->next) - symbol_type_set (list->content.sym, $2, @2); + symbol_type_set (list->content.sym, $2, @2); symbol_list_free ($3); } ; precedence_declaration: - precedence_declarator type.opt symbols.prec + precedence_declarator tag.opt symbols.prec { symbol_list *list; ++current_prec; for (list = $3; list; list = list->next) - { - symbol_type_set (list->content.sym, current_type, @2); - symbol_precedence_set (list->content.sym, current_prec, $1, @1); - } + { + symbol_type_set (list->content.sym, current_type, @2); + symbol_precedence_set (list->content.sym, current_prec, $1, @1); + } symbol_list_free ($3); current_type = NULL; } ; precedence_declarator: - "%left" { $$ = left_assoc; } -| "%right" { $$ = right_assoc; } -| "%nonassoc" { $$ = non_assoc; } + "%left" { $$ = left_assoc; } +| "%right" { $$ = right_assoc; } +| "%nonassoc" { $$ = non_assoc; } +| "%precedence" { $$ = precedence_assoc; } ; -type.opt: +tag.opt: /* Nothing. */ { current_type = NULL; } -| TYPE { current_type = $1; tag_seen = true; } +| TAG { current_type = $1; tag_seen = true; } ; /* Just like symbols.1 but accept INT for the sake of POSIX. */ @@ -460,9 +507,9 @@ symbols.prec: ; symbol.prec: - symbol { $$ = $1; } - | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); } - ; + symbol { $$ = $1; } +| symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); } +; /* One or more symbols to be %typed. */ symbols.1: @@ -478,15 +525,19 @@ generic_symlist: ; generic_symlist_item: - symbol { $$ = symbol_list_sym_new ($1, @1); } -| TYPE { $$ = symbol_list_type_new ($1, @1); } -| "<*>" { $$ = symbol_list_default_tagged_new (@1); } -| "<>" { $$ = symbol_list_default_tagless_new (@1); } + symbol { $$ = symbol_list_sym_new ($1, @1); } +| tag { $$ = symbol_list_type_new ($1, @1); } +; + +tag: + TAG +| "<*>" { $$ = uniqstr_new ("*"); } +| "<>" { $$ = uniqstr_new (""); } ; /* One token definition. */ symbol_def: - TYPE + TAG { current_type = $1; tag_seen = true; @@ -524,9 +575,9 @@ symbol_defs.1: ; - /*------------------------------------------. - | The grammar section: between the two %%. | - `------------------------------------------*/ + /*------------------------------------------. + | The grammar section: between the two %%. | + `------------------------------------------*/ grammar: rules_or_grammar_declaration @@ -561,16 +612,18 @@ rhses.1: rhs: /* Nothing. */ { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location, - current_lhs_named_ref); } + current_lhs_named_ref); } | rhs symbol named_ref.opt { grammar_current_rule_symbol_append ($2, @2, $3); } | rhs "{...}" named_ref.opt - { grammar_current_rule_action_append ($2, @2, $3); } + { grammar_current_rule_action_append ($2, @2, $3, false); } +| rhs "%?{...}" + { grammar_current_rule_action_append ($2, @2, NULL, true); } | rhs "%prec" symbol { grammar_current_rule_prec_set ($3, @3); } | rhs "%dprec" INT { grammar_current_rule_dprec_set ($3, @3); } -| rhs "%merge" TYPE +| rhs "%merge" TAG { grammar_current_rule_merge_set ($3, @3); } ; @@ -580,10 +633,9 @@ named_ref.opt: BRACKETED_ID { $$ = named_ref_new($1, @1); } ; - -/*----------------------------* - | variable and content.opt. | - *---------------------------*/ +/*---------------------------. +| variable and content.opt. | +`---------------------------*/ /* The STRING form of variable is deprecated and is not M4-friendly. For example, M4 fails for `%define "[" "value"'. */ @@ -600,9 +652,9 @@ content.opt: ; -/*-------------* - | braceless. | - *-------------*/ +/*------------. +| braceless. | +`------------*/ braceless: "{...}" @@ -617,9 +669,9 @@ braceless: ; -/*---------------* - | Identifiers. | - *---------------*/ +/*--------------. +| Identifiers. | +`--------------*/ /* Identifiers are returned as uniqstr values by the scanner. Depending on their use, we may need to make them genuine symbols. */ @@ -669,7 +721,6 @@ epilogue.opt: %% - /* Return the location of the left-hand side of a rule whose right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in the right-hand side, and return an empty location equal to the end @@ -692,49 +743,53 @@ lloc_default (YYLTYPE const *rhs, int n) for (i = 1; i <= n; i++) if (! equal_boundaries (rhs[i].start, rhs[i].end)) { - loc.start = rhs[i].start; - break; + loc.start = rhs[i].start; + break; } return loc; } -/* Add a lex-param or a parse-param (depending on TYPE) with - declaration DECL and location LOC. */ - static void -add_param (char const *type, char *decl, location loc) +add_param (param_type type, char *decl, location loc) { static char const alphanum[26 + 26 + 1 + 10] = "abcdefghijklmnopqrstuvwxyz" "ABCDEFGHIJKLMNOPQRSTUVWXYZ" "_" "0123456789"; + char const *name_start = NULL; - char *p; - - /* Stop on last actual character. */ - for (p = decl; p[1]; p++) - if ((p == decl - || ! memchr (alphanum, p[-1], sizeof alphanum)) - && memchr (alphanum, p[0], sizeof alphanum - 10)) - name_start = p; - - /* Strip the surrounding '{' and '}', and any blanks just inside - the braces. */ - while (*--p == ' ' || *p == '\t') - continue; - p[1] = '\0'; - while (*++decl == ' ' || *decl == '\t') - continue; + { + char *p; + /* Stop on last actual character. */ + for (p = decl; p[1]; p++) + if ((p == decl + || ! memchr (alphanum, p[-1], sizeof alphanum)) + && memchr (alphanum, p[0], sizeof alphanum - 10)) + name_start = p; + + /* Strip the surrounding '{' and '}', and any blanks just inside + the braces. */ + --p; + while (c_isspace ((unsigned char) *p)) + --p; + p[1] = '\0'; + ++decl; + while (c_isspace ((unsigned char) *decl)) + ++decl; + } if (! name_start) - complain_at (loc, _("missing identifier in parameter declaration")); + complain (&loc, complaint, _("missing identifier in parameter declaration")); else { char *name = xmemdup0 (name_start, strspn (name_start, alphanum)); - muscle_pair_list_grow (type, decl, name); + if (type & param_lex) + muscle_pair_list_grow ("lex_param", decl, name); + if (type & param_parse) + muscle_pair_list_grow ("parse_param", decl, name); free (name); } @@ -747,8 +802,8 @@ version_check (location const *loc, char const *version) { if (strverscmp (version, PACKAGE_VERSION) > 0) { - complain_at (*loc, "require bison %s, but have %s", - version, PACKAGE_VERSION); + complain (loc, complaint, "require bison %s, but have %s", + version, PACKAGE_VERSION); exit (EX_MISMATCH); } } @@ -756,7 +811,7 @@ version_check (location const *loc, char const *version) static void gram_error (location const *loc, char const *msg) { - complain_at (*loc, "%s", msg); + complain (loc, complaint, "%s", msg); } char const *