X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/e9955c83734d0a545d7822a1feb9c4a8038a62cb..ba29c4ec2590be025f37df6c722ad4a9a16f78c2:/src/scan-gram.l diff --git a/src/scan-gram.l b/src/scan-gram.l index e6abed5f..02d47e5a 100644 --- a/src/scan-gram.l +++ b/src/scan-gram.l @@ -31,24 +31,58 @@ #include "reader.h" /* Each time we match a string, move the end cursor to its end. */ +#define YY_USER_INIT \ +do { \ + LOCATION_RESET (*yylloc); \ + /* This is only to avoid GCC warnings. */ \ + if (yycontrol) {;}; \ +} while (0) + #define YY_USER_ACTION LOCATION_COLUMNS (*yylloc, yyleng) #define YY_LINES LOCATION_LINES (*yylloc, yyleng); lineno += yyleng; #define YY_STEP LOCATION_STEP (*yylloc) -/* Appending to the STRING_OBSTACK. */ -#define YY_INIT obstack_init (&string_obstack) -#define YY_GROW obstack_grow (&string_obstack, yytext, yyleng) -#define YY_FINISH obstack_1grow (&string_obstack, '\0'); yylval->string = obstack_finish (&string_obstack); -/* This is only to avoid GCC warnings. */ -#define YY_USER_INIT if (yycontrol) {;}; +/* 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 braces_level = 0; static int percent_percent_count = 0; -static void handle_dollar PARAMS ((char *cp)); -static void handle_at PARAMS ((char *cp)); +static void handle_dollar PARAMS ((braced_code_t code_kind, + char *cp, location_t location)); +static void handle_at PARAMS ((braced_code_t code_kind, + char *cp, location_t location)); %} %x SC_COMMENT @@ -67,7 +101,7 @@ blanks [ \t\f]+ start of the next token. */ #define TR_POS 0 #if TR_POS - fprintf (stderr, "FOO1: "); + fprintf (stderr, "FOO1: %p: ", yylloc); LOCATION_PRINT (stderr, *yylloc); fprintf (stderr, "\n"); #endif @@ -89,6 +123,7 @@ blanks [ \t\f]+ "%debug" return PERCENT_DEBUG; "%define" return PERCENT_DEFINE; "%defines" return PERCENT_DEFINES; + "%destructor" return PERCENT_DESTRUCTOR; "%error"[-_]"verbose" return PERCENT_ERROR_VERBOSE; "%expect" return PERCENT_EXPECT; "%file-prefix" return PERCENT_FILE_PREFIX; @@ -101,6 +136,7 @@ blanks [ \t\f]+ "%nterm" return PERCENT_NTERM; "%output" return PERCENT_OUTPUT; "%prec" return PERCENT_PREC; + "%printer" return PERCENT_PRINTER; "%pure"[-_]"parser" return PERCENT_PURE_PARSER; "%right" return PERCENT_RIGHT; "%skeleton" return PERCENT_SKELETON; @@ -121,31 +157,36 @@ blanks [ \t\f]+ {eols} YY_LINES; YY_STEP; {blanks} YY_STEP; {id} { - YY_INIT; YY_GROW; YY_FINISH; - yylval->symbol = getsym (yylval->string); + yylval->symbol = getsym (yytext, *yylloc); return ID; } {int} yylval->integer = strtol (yytext, 0, 10); return INT; /* Characters. We don't check there is only one. */ - \' YY_INIT; YY_GROW; yy_push_state (SC_ESCAPED_CHARACTER); + \' YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER); /* Strings. */ - \" YY_INIT; YY_GROW; yy_push_state (SC_ESCAPED_STRING); + \" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING); /* Comments. */ "/*" yy_push_state (SC_COMMENT); "//".* YY_STEP; /* Prologue. */ - "%{" YY_INIT; yy_push_state (SC_PROLOGUE); + "%{" yy_push_state (SC_PROLOGUE); /* Code in between braces. */ - "{" YY_INIT; YY_GROW; ++braces_level; yy_push_state (SC_BRACED_CODE); + "{" YY_OBS_GROW; ++braces_level; yy_push_state (SC_BRACED_CODE); /* A type. */ - "<"[^>]+">" YY_INIT; obstack_grow (&string_obstack, yytext + 1, yyleng - 2); YY_FINISH; return TYPE; + "<"[^>]+">" { + obstack_grow (&string_obstack, yytext + 1, yyleng - 2); + YY_OBS_FINISH; + yylval->string = last_string; + return TYPE; + } + "%%" { if (++percent_percent_count == 2) @@ -169,8 +210,8 @@ blanks [ \t\f]+ { - \[ obstack_sgrow (&string_obstack, "@<:@"); - \] obstack_sgrow (&string_obstack, "@:>@"); + \[ if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@<:@"); + \] if (YY_START != SC_COMMENT) obstack_sgrow (&string_obstack, "@:>@"); } @@ -188,14 +229,14 @@ blanks [ \t\f]+ } else { - YY_GROW; + YY_OBS_GROW; } yy_pop_state (); } - [^\[\]*\n\r]+ if (yy_top_state () != INITIAL) YY_GROW; - {eols} if (yy_top_state () != INITIAL) YY_GROW; YY_LINES; - . /* Stray `*'. */if (yy_top_state () != INITIAL) YY_GROW; + [^\[\]*\n\r]+ if (yy_top_state () != INITIAL) YY_OBS_GROW; + {eols} if (yy_top_state () != INITIAL) YY_OBS_GROW; YY_LINES; + . /* Stray `*'. */if (yy_top_state () != INITIAL) YY_OBS_GROW; <> { LOCATION_PRINT (stderr, *yylloc); @@ -214,13 +255,14 @@ blanks [ \t\f]+ { \" { assert (yy_top_state () == INITIAL); - YY_GROW; - YY_FINISH; + YY_OBS_GROW; + YY_OBS_FINISH; + yylval->string = last_string; yy_pop_state (); return STRING; } - [^\"\n\r\\]+ YY_GROW; + [^\"\n\r\\]+ YY_OBS_GROW; {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES; @@ -228,7 +270,8 @@ blanks [ \t\f]+ LOCATION_PRINT (stderr, *yylloc); fprintf (stderr, ": unexpected end of file in a string\n"); assert (yy_top_state () == INITIAL); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; yy_pop_state (); return STRING; } @@ -242,21 +285,20 @@ blanks [ \t\f]+ { \' { - YY_GROW; + YY_OBS_GROW; assert (yy_top_state () == INITIAL); { - char c; - YY_FINISH; - c = yylval->string[1]; - yylval->symbol = getsym (yylval->string); - symbol_class_set (yylval->symbol, token_sym); - symbol_user_token_number_set (yylval->symbol, (unsigned int) c); + YY_OBS_FINISH; + yylval->symbol = getsym (last_string, *yylloc); + symbol_class_set (yylval->symbol, token_sym, *yylloc); + symbol_user_token_number_set (yylval->symbol, last_string[1], *yylloc); + YY_OBS_FREE; yy_pop_state (); return ID; } } - [^\'\n\r\\] YY_GROW; + [^\'\n\r\\] YY_OBS_GROW; {eols} obstack_1grow (&string_obstack, '\n'); YY_LINES; @@ -264,7 +306,8 @@ blanks [ \t\f]+ LOCATION_PRINT (stderr, *yylloc); fprintf (stderr, ": unexpected end of file in a character\n"); assert (yy_top_state () == INITIAL); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; yy_pop_state (); return CHARACTER; } @@ -282,7 +325,7 @@ blanks [ \t\f]+ if (c > 255) { LOCATION_PRINT (stderr, *yylloc); - fprintf (stderr, ": invalid escape: %s\n", yytext); + fprintf (stderr, ": invalid escape: %s\n", quote (yytext)); YY_STEP; } else @@ -301,11 +344,13 @@ blanks [ \t\f]+ \\t obstack_1grow (&string_obstack, '\t'); \\v obstack_1grow (&string_obstack, '\v'); \\[\\""] obstack_1grow (&string_obstack, yytext[1]); - \\. { + \\(.|\n) { LOCATION_PRINT (stderr, *yylloc); - fprintf (stderr, ": unrecognized escape: %s\n", yytext); - YY_GROW; + fprintf (stderr, ": unrecognized escape: %s\n", quote (yytext)); + YY_OBS_GROW; } + /* FLex wants this rule, in case of a `\<>'. */ + \\ YY_OBS_GROW; } @@ -317,15 +362,17 @@ blanks [ \t\f]+ { \' { - YY_GROW; + YY_OBS_GROW; assert (yy_top_state () != INITIAL); yy_pop_state (); } - [^\[\]\'\n\r\\] YY_GROW; - \\. YY_GROW; + [^\[\]\'\n\r\\]+ YY_OBS_GROW; + \\(.|\n) YY_OBS_GROW; + /* FLex wants this rule, in case of a `\<>'. */ + \\ YY_OBS_GROW; - {eols} YY_GROW; YY_LINES; + {eols} YY_OBS_GROW; YY_LINES; <> { LOCATION_PRINT (stderr, *yylloc); @@ -345,14 +392,16 @@ blanks [ \t\f]+ { \" { assert (yy_top_state () != INITIAL); - YY_GROW; + YY_OBS_GROW; yy_pop_state (); } - [^\[\]\"\n\r\\]+ YY_GROW; - \\. YY_GROW; + [^\[\]\"\n\r\\]+ YY_OBS_GROW; + \\(.|\n) YY_OBS_GROW; + /* FLex wants this rule, in case of a `\<>'. */ + \\ YY_OBS_GROW; - {eols} YY_GROW; YY_LINES; + {eols} YY_OBS_GROW; YY_LINES; <> { LOCATION_PRINT (stderr, *yylloc); @@ -370,14 +419,17 @@ blanks [ \t\f]+ { /* Characters. We don't check there is only one. */ - \' YY_GROW; yy_push_state (SC_CHARACTER); + \' YY_OBS_GROW; yy_push_state (SC_CHARACTER); /* Strings. */ - \" YY_GROW; yy_push_state (SC_STRING); + \" YY_OBS_GROW; yy_push_state (SC_STRING); /* Comments. */ - "/*" YY_GROW; yy_push_state (SC_COMMENT); - "//".* YY_GROW; + "/*" YY_OBS_GROW; yy_push_state (SC_COMMENT); + "//".* YY_OBS_GROW; + + /* Not comments. */ + "/" YY_OBS_GROW; } @@ -389,32 +441,36 @@ blanks [ \t\f]+ { "}" { - YY_GROW; + YY_OBS_GROW; if (--braces_level == 0) { yy_pop_state (); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; return BRACED_CODE; } } - "{" YY_GROW; braces_level++; + "{" YY_OBS_GROW; braces_level++; - "$"("<".*">")?(-?[0-9]+|"$") { handle_dollar (yytext); } - "@"(-?[0-9]+|"$") { handle_at (yytext); } + "$"("<"[^>]+">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code, + yytext, *yylloc); } + "@"(-?[0-9]+|"$") { handle_at (current_braced_code, + yytext, *yylloc); } - [^\[\]$/\'\"@\{\}\n\r]+ YY_GROW; - {eols} YY_GROW; YY_LINES; + [^$@\[\]/\'\"\{\}\n\r]+ YY_OBS_GROW; + {eols} YY_OBS_GROW; YY_LINES; /* A lose $, or /, or etc. */ - . YY_GROW; + . YY_OBS_GROW; <> { LOCATION_PRINT (stderr, *yylloc); fprintf (stderr, ": unexpected end of file in a braced code\n"); yy_pop_state (); - YY_FINISH; - return PROLOGUE; + YY_OBS_FINISH; + yylval->string = last_string; + return BRACED_CODE; } } @@ -428,22 +484,23 @@ blanks [ \t\f]+ { "%}" { yy_pop_state (); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; return PROLOGUE; } - [^\[\]%\n\r]+ YY_GROW; - "%"+[^%\}\n\r]+ YY_GROW; - {eols} YY_GROW; YY_LINES; + [^%\[\]/\'\"\n\r]+ YY_OBS_GROW; + "%" YY_OBS_GROW; + {eols} YY_OBS_GROW; YY_LINES; <> { LOCATION_PRINT (stderr, *yylloc); fprintf (stderr, ": unexpected end of file in a prologue\n"); yy_pop_state (); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; return PROLOGUE; } - } @@ -454,11 +511,12 @@ blanks [ \t\f]+ { - ([^\[\]]|{eols})+ YY_GROW; + ([^\[\]]|{eols})+ YY_OBS_GROW; <> { yy_pop_state (); - YY_FINISH; + YY_OBS_FINISH; + yylval->string = last_string; return EPILOGUE; } } @@ -467,28 +525,18 @@ blanks [ \t\f]+ %% /*------------------------------------------------------------------. -| CP is pointing to a wannabee semantic value (i.e., a `$'). | +| TEXT is pointing to a wannabee semantic value (i.e., a `$'). | | | | Possible inputs: $[]($|integer) | | | | Output to the STRING_OBSTACK a reference to this semantic value. | `------------------------------------------------------------------*/ -static void -handle_dollar (char *cp) +static inline void +handle_action_dollar (char *text, location_t location) { const char *type_name = NULL; - - /* RULE_LENGTH is the number of values in the current rule so far, - which says where to find `$0' with respect to the top of the - stack. It is not the same as the rule->length in the case of mid - rule actions. */ - int rule_length = 0; - symbol_list *rhs; - for (rhs = current_rule->next; rhs; rhs = rhs->next) - ++rule_length; - - ++cp; + char *cp = text + 1; /* Get the type name if explicit. */ if (*cp == '<') @@ -503,10 +551,10 @@ handle_dollar (char *cp) if (*cp == '$') { if (!type_name) - type_name = get_type_name (0, current_rule); + type_name = symbol_list_n_type_name_get (current_rule, location, 0); if (!type_name && typed) - complain (_("$$ of `%s' has no declared type"), - current_rule->sym->tag); + complain_at (location, _("$$ of `%s' has no declared type"), + symbol_tag_get (current_rule->sym)); if (!type_name) type_name = ""; obstack_fgrow1 (&string_obstack, @@ -514,17 +562,23 @@ handle_dollar (char *cp) } else if (isdigit (*cp) || *cp == '-') { + /* RULE_LENGTH is the number of values in the current rule so + far, which says where to find `$0' with respect to the top of + the stack. It is not the same as the rule->length in the + case of mid rule actions. */ + int rule_length = symbol_list_length (current_rule->next); int n = strtol (cp, &cp, 10); if (n > rule_length) - complain (_("invalid value: %s%d"), "$", n); + complain_at (location, _("invalid value: %s%d"), "$", n); else { if (!type_name && n > 0) - type_name = get_type_name (n, current_rule); + type_name = symbol_list_n_type_name_get (current_rule, location, + n); if (!type_name && typed) - complain (_("$%d of `%s' has no declared type"), - n, current_rule->sym->tag); + complain_at (location, _("$%d of `%s' has no declared type"), + n, symbol_tag_get (current_rule->sym)); if (!type_name) type_name = ""; obstack_fgrow3 (&string_obstack, @@ -534,31 +588,60 @@ handle_dollar (char *cp) } else { - char buf[] = "$c"; - buf[1] = *cp; - complain (_("%s is invalid"), quote (buf)); + complain_at (location, _("%s is invalid"), quote (text)); } } -/*-------------------------------------------------------. -| CP is pointing to a location (i.e., a `@'). Output to | -| STRING_OBSTACK a reference to this location. | -`-------------------------------------------------------*/ + +/*---------------------------------------------------------------. +| TEXT is expexted tp 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, _("%s is invalid"), quote (text)); +} + + +/*-----------------------------------------------------------------. +| Dispatch onto handle_action_dollar, or handle_destructor_dollar, | +| depending upon CODE_KIND. | +`-----------------------------------------------------------------*/ static void -handle_at (char *cp) +handle_dollar (braced_code_t braced_code_kind, + char *text, location_t location) { - /* RULE_LENGTH is the number of values in the current rule so far, - which says where to find `$0' with respect to the top of the - stack. It is not the same as the rule->length in the case of mid - rule actions. */ - int rule_length = 0; - symbol_list *rhs; - for (rhs = current_rule->next; rhs; rhs = rhs->next) - ++rule_length; + switch (braced_code_kind) + { + case action_braced_code: + handle_action_dollar (text, location); + break; + + case destructor_braced_code: + case printer_braced_code: + handle_symbol_code_dollar (text, location); + break; + } +} + +/*------------------------------------------------------. +| TEXT is a location token (i.e., a `@...'). Output to | +| STRING_OBSTACK a reference to this location. | +`------------------------------------------------------*/ + +static inline void +handle_action_at (char *text, location_t location) +{ + char *cp = text + 1; locations_flag = 1; - ++cp; if (*cp == '$') { @@ -566,17 +649,82 @@ handle_at (char *cp) } else if (isdigit (*cp) || *cp == '-') { + /* RULE_LENGTH is the number of values in the current rule so + far, which says where to find `$0' with respect to the top of + the stack. It is not the same as the rule->length in the + case of mid rule actions. */ + int rule_length = symbol_list_length (current_rule->next); int n = strtol (cp, &cp, 10); + if (n > rule_length) - complain (_("invalid value: %s%d"), "@", n); + complain_at (location, _("invalid value: %s%d"), "@", n); else obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[", rule_length, n); } else { - char buf[] = "@c"; - buf[1] = *cp; - complain (_("%s is invalid"), quote (buf)); + complain_at (location, _("%s is invalid"), quote (text)); } } + + +/*---------------------------------------------------------------. +| TEXT is expexted tp be @$ 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, _("%s is invalid"), 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; + } +} + + +/*-------------------------. +| 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); +}