--- parser3/src/main/compile.y 2001/03/06 14:28:35 1.54 +++ parser3/src/main/compile.y 2001/03/27 14:51:29 1.119 @@ -1,28 +1,47 @@ -/* - $Id: compile.y,v 1.54 2001/03/06 14:28:35 paf Exp $ +/** @file + Parser: compiler(lexical parser and grammar). + + Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) + + Author: Alexander Petrosyan (http://design.ru/paf) + + $Id: compile.y,v 1.119 2001/03/27 14:51:29 paf Exp $ +*/ + +/** + @todo parser4: + - cache compiled code from request to request. to do that... + -#: make method definitions, @CLASS, @BASE, @USE instructions, + which would be executed afterwards, and actions + now performed at compile time would be delayed to run time. + -#: make cache expiration on time and on disk-change of class source + -#: in apache use subpools for compiled class storage + -#: in iis make up specialized Pool object for that */ %{ -#define YYSTYPE Array/**/ * +#define YYSTYPE Array/**/ * #define YYPARSE_PARAM pc #define YYLEX_PARAM pc #define YYDEBUG 1 -#define YYERROR_VERBOSE +#define YYERROR_VERBOSE 1 #define yyerror(msg) real_yyerror((parse_control *)pc, msg) #define YYPRINT(file, type, value) yyprint(file, type, value) +#include "pa_config_includes.h" #include -#include #include +#include #include "compile_tools.h" #include "pa_value.h" #include "pa_request.h" #include "pa_vobject.h" #include "pa_vdouble.h" +#include "pa_globals.h" -#define SELF_NAME "self" -#define USES_NAME "USES" +#define SELF_ELEMENT_NAME "self" +#define USE_CONTROL_METHOD_NAME "USE" int real_yyerror(parse_control *pc, char *s); static void yyprint(FILE *file, int type, YYSTYPE value); @@ -30,10 +49,11 @@ int yylex(YYSTYPE *lvalp, void *pc); // local convinient inplace typecast & var -#define PC ((parse_control *)pc) -#define POOL *PC->pool +#define PC (*(parse_control *)pc) +#define POOL (*PC.pool) #undef NEW #define NEW new(POOL) +#ifndef DOXYGEN %} %pure_parser @@ -41,18 +61,61 @@ int yylex(YYSTYPE *lvalp, void *pc); %token EON %token STRING %token BOGUS -%token LOGICAL_AND LOGICAL_OR -%% +%token BAD_STRING_COMPARISON_OPERATOR +%token BAD_HEX_LITERAL + +%token LAND "&&" +%token LOR "||" +%token LXOR "##" + +%token NLE "<=" +%token NGE ">=" +%token NEQ "==" +%token NNE "!=" + +%token SLT "lt" +%token SGT "gt" +%token SLE "le" +%token SGE "ge" +%token SEQ "eq" +%token SNE "ne" + +%token DEF "def" +%token IN "in" +%token FEXISTS "-f" +%token IS "is" + +/* logical */ +%left "is" +%left "lt" "gt" "le" "ge" +%left "eq" "ne" +%left '<' '>' "<=" ">=" "##" +%left "==" "!=" +%left "||" +%left "&&" +%left "def" "in" "-f" +%left '!' + +/* bitwise */ +%left '#' +%left '&' '|' +%left '~' + +/* numerical */ +%left '-' '+' +%left '*' '/' '%' +%left NEG /* negation: unary - */ +%% all: one_big_piece { - String& name_main=*NEW String(POOL); - name_main.APPEND_CONST(MAIN_METHOD_NAME); - Array& param_names=*NEW Array(POOL); - Array& local_names=*NEW Array(POOL); - Method& method=*NEW Method(POOL, name_main, param_names, local_names, *$1); - PC->vclass->add_method(name_main, method); + Method& method=*NEW Method(POOL, + *main_method_name, + 0, 0, /*min, max numbered_params_count*/ + 0/*param_names*/, 0/*local_names*/, + $1/*parser_code*/, 0/*native_code*/); + PC.cclass->add_method(*main_method_name, method); } | methods; @@ -63,47 +126,71 @@ method: control_method | code_method; control_method: '@' STRING '\n' control_strings { - String& name=*SLA2S($2); + const String& command=*SLA2S($2); YYSTYPE strings_code=$4; if(strings_code->size()<1*2) { - strcpy(PC->error, "@"); - strcat(PC->error, name.cstr()); - strcat(PC->error, " is empty"); + strcpy(PC.error, "@"); + strcat(PC.error, command.cstr()); + strcat(PC.error, " is empty"); YYERROR; } - if(name==CLASS_NAME) { - if(strings_code->size()==1*2) - PC->vclass->set_name(*SLA2S(strings_code)); - else { - strcpy(PC->error, "@"CLASS_NAME" must contain sole name"); + if(command==CLASS_NAME) { + if(PC.cclass!=&PC.request->ROOT) { // already changed from default? + strcpy(PC.error, "class already have a name '"); + strncat(PC.error, PC.cclass->name().cstr(), 100); + strcat(PC.error, "'"); + YYERROR; + } + if(strings_code->size()==1*2) { + // new class' name + const String *name=SLA2S(strings_code); + // creating the class + PC.cclass=NEW VClass(POOL); + PC.cclass->set_name(*name); + // defaulting base. may change with @BASE + PC.cclass->set_base(PC.request->ROOT); + // append to request's classes + PC.request->classes().put(*name, PC.cclass); + } else { + strcpy(PC.error, "@"CLASS_NAME" must contain sole name"); YYERROR; } } else { - if(name==USES_NAME) { - for(int i=0; isize(); i+=2) { - String *file=SLA2S(strings_code, i); - file->APPEND_CONST(".p"); - PC->request->use(file->cstr(), 0); + if(command==USE_CONTROL_METHOD_NAME) { + for(int i=0; isize(); i+=2) + PC.request->use_file( + PC.request->absolute(*SLA2S(strings_code, i))); + } else if(command==BASE_NAME) { + if(PC.cclass->base()!=&PC.request->ROOT) { // already changed from default? + strcpy(PC.error, "class already have a base '"); + strncat(PC.error, PC.cclass->base()->name().cstr(), 100); + strcat(PC.error, "'"); + YYERROR; } - } else if(name==BASE_NAME) { if(strings_code->size()==1*2) { - String& base_name=*SLA2S(strings_code); + const String& base_name=*SLA2S(strings_code); VClass *base=static_cast( - PC->request->classes().get(base_name)); + PC.request->classes().get(base_name)); if(!base) { - strcpy(PC->error, base_name.cstr()); - strcat(PC->error, ": undefined class in @"BASE_NAME); + strcpy(PC.error, base_name.cstr()); + strcat(PC.error, ": undefined class in @"BASE_NAME); + YYERROR; + } + // @CLASS == @BASE sanity check + if(PC.cclass==base) { + strcpy(PC.error, "@"CLASS_NAME" equals @"BASE_NAME); YYERROR; } - PC->vclass->set_base(*base); + PC.cclass->set_base(*base); } else { - strcpy(PC->error, "@"BASE_NAME" must contain sole name"); + strcpy(PC.error, "@"BASE_NAME" must contain sole name"); YYERROR; } } else { - strcpy(PC->error, name.cstr()); - strcat(PC->error, ": invalid special name. valid names are " - CLASS_NAME", "USES_NAME" and "BASE_NAME); + strcpy(PC.error, "'"); + strncat(PC.error, command.cstr(), MAX_STRING/2); + strcat(PC.error, "' invalid special name. valid names are " + "'"CLASS_NAME"', '"USE_CONTROL_METHOD_NAME"' and '"BASE_NAME"'"); YYERROR; } } @@ -117,17 +204,27 @@ code_method: '@' STRING bracketed_maybe_ const String *name=SLA2S($2); YYSTYPE params_names_code=$3; - Array& params_names=*NEW Array(POOL); - for(int i=0; isize(); i+=2) - params_names+=SLA2S(params_names_code, i); + Array *params_names=0; + if(int size=params_names_code->size()) { + params_names=NEW Array(POOL); + for(int i=0; isize(); i+=2) - locals_names+=SLA2S(locals_names_code, i); + Array *locals_names=0; + if(int size=locals_names_code->size()) { + locals_names=NEW Array(POOL); + for(int i=0; ivclass->add_method(*name, method); + Method& method=*NEW Method(POOL, + *name, + 0, 0/*min,max numbered_params_count*/, + params_names, locals_names, + $7, 0); + PC.cclass->add_method(*name, method); }; maybe_bracketed_strings: empty | bracketed_maybe_strings; @@ -141,20 +238,18 @@ maybe_comment: empty | STRING; maybe_codes: empty | codes; -codes: code | codes code { - $$=$1; - P($$, $2); -}; -code: write_str_literal | action; +codes: code | codes code { $$=$1; P($$, $2) }; +code: write_string | action; action: get | put | with | call; /* get */ -get: '$' get_name { - $$=$2; /* stack: resulting value */ - O($$, OP_WRITE); /* value=pop; write(value) */ +get: get_value { + $$=$1; /* stack: resulting value */ + O($$, OP_WRITE_VALUE); /* value=pop; wcontext.write(value) */ }; -get_name: name_without_curly_rdive EON | name_in_curly_rdive; +get_value: '$' get_name_value { $$=$2 } +get_name_value: name_without_curly_rdive EON | name_in_curly_rdive; name_in_curly_rdive: '{' name_without_curly_rdive '}' { $$=$2 }; name_without_curly_rdive: name_without_curly_rdive_read @@ -163,8 +258,8 @@ name_without_curly_rdive: name_without_curly_rdive_read: name_without_curly_rdive_code { $$=N(POOL); Array *diving_code=$1; - String *first_name=SLA2S(diving_code); - if(first_name && *first_name==SELF_NAME) { + const String *first_name=SLA2S(diving_code); + if(first_name && *first_name==SELF_ELEMENT_NAME) { O($$, OP_WITH_SELF); /* stack: starting context */ P($$, diving_code, /* skip over... */ @@ -185,10 +280,9 @@ name_without_curly_rdive_code: name_adva /* put */ -put: '$' name_expr_wdive constructor_value { +put: '$' name_expr_wdive construct { $$=$2; /* stack: context,name */ P($$, $3); /* stack: context,name,constructor_value */ - O($$, OP_CONSTRUCT); /* value=pop; name=pop; context=pop; construct(context,name,value) */ }; name_expr_wdive: name_expr_wdive_write @@ -197,8 +291,8 @@ name_expr_wdive: name_expr_wdive_write: name_expr_dive_code { $$=N(POOL); Array *diving_code=$1; - String *first_name=SLA2S(diving_code); - if(first_name && *first_name==SELF_NAME) { + const String *first_name=SLA2S(diving_code); + if(first_name && *first_name==SELF_ELEMENT_NAME) { O($$, OP_WITH_SELF); /* stack: starting context */ P($$, diving_code, /* skip over... */ @@ -216,58 +310,93 @@ name_expr_wdive_root: ':' name_expr_dive }; name_expr_wdive_class: class_prefix name_expr_dive_code { $$=$1; P($$, $2) }; -constructor_value: - '[' constructor_code_value ']' { $$=$2 } -| '(' expression_value ')' { $$=$2 } +construct: construct_by_code | construct_by_expr; +construct_by_code: '[' any_constructor_code_value ']' { + $$=$2; /* stack: context, name, value */ + O($$, OP_CONSTRUCT_VALUE); /* value=pop; name=pop; context=pop; construct(context,name,value) */ +} +; +construct_by_expr: '(' expr_value ')' { + $$=$2; /* stack: context, name, value */ + O($$, OP_CONSTRUCT_EXPR); /* value=pop; name=pop; context=pop; construct(context,name,value) */ +} ; -constructor_code_value: - empty_value /* optimized $var[] case */ +any_constructor_code_value: + empty_string_value /* optimized $var[] case */ | STRING /* optimized $var[STRING] case */ -| complex_constructor_code_value /* $var[something complex] */ +| constructor_code_value /* $var[something complex] */ ; -complex_constructor_code_value: complex_constructor_code { +constructor_code_value: constructor_code { $$=N(POOL); O($$, OP_CREATE_EWPOOL); /* stack: empty write context */ - P($$, $1); /* some codes to that context */ + P($$, $1); /* some code that writes to that context */ O($$, OP_REDUCE_EWPOOL); /* context=pop; stack: context.value() */ }; -complex_constructor_code: codes__excluding_sole_str_literal; +constructor_code: codes__excluding_sole_str_literal; codes__excluding_sole_str_literal: action | code codes { $$=$1; P($$, $2) }; /* call */ -call: '^' call_name store_params EON { /* ^field.$method{vasya} */ +call: call_value { + $$=$1; /* stack: value */ + O($$, OP_WRITE_VALUE); /* value=pop; wcontext.write(value) */ +}; +call_value: '^' call_name store_params EON { /* ^field.$method{vasya} */ $$=$2; /* with_xxx,diving code; stack: context,method_junction */ O($$, OP_GET_METHOD_FRAME); /* stack: context,method_frame */ P($$, $3); /* filling method_frame.store_params */ - O($$, OP_CALL); /* method_frame=pop; ncontext=pop; call(ncontext,method_frame) */ + O($$, OP_CALL); /* method_frame=pop; ncontext=pop; call(ncontext,method_frame) stack: value */ }; call_name: name_without_curly_rdive; store_params: store_param | store_params store_param { $$=$1; P($$, $2) }; -store_param: store_round_param | store_curly_param; -store_round_param: '[' store_param_parts ']' {$$=$2}; -store_param_parts: - store_param_part -| store_param_parts ';' store_param_part { $$=$1; P($$, $3) } +store_param: + store_square_param +| store_round_param +| store_curly_param ; -store_curly_param: '{' maybe_codes '}' { - $$=N(POOL); - PCA($$, $2); - O($$, OP_STORE_PARAM); -}; -store_param_part: - empty /* optimized () case */ -| STRING { /* optimized (STRING) case */ +store_square_param: '[' store_code_param_parts ']' {$$=$2}; +store_round_param: '(' store_expr_param_parts ')' {$$=$2}; +store_curly_param: '{' store_curly_param_parts '}' {$$=$2}; +store_code_param_parts: + empty /* optimized [] case */ +| store_code_param_parts_not_empty +; +store_code_param_parts_not_empty: + store_code_param_part +| store_code_param_parts ';' store_code_param_part { $$=$1; P($$, $3) } +; +store_expr_param_parts: + store_expr_param_part +| store_expr_param_parts ';' store_expr_param_part { $$=$1; P($$, $3) } +; +store_curly_param_parts: + store_curly_param_part +| store_curly_param_parts ';' store_curly_param_part { $$=$1; P($$, $3) } +; +store_code_param_part: + STRING { /* optimized [STRING] case */ $$=$1; O($$, OP_STORE_PARAM); } -| complex_constructor_code_value { /* (something complex) */ +| constructor_code_value { /* [something complex] */ $$=$1; O($$, OP_STORE_PARAM); } ; +store_expr_param_part: write_expr_value { + $$=N(POOL); + PEA($$, $1); +}; +store_curly_param_part: maybe_codes { + $$=N(POOL); + PCA($$, $1); +}; +write_expr_value: expr_value { + $$=$1; + O($$, OP_WRITE_EXPR_RESULT); +}; /* name */ @@ -300,7 +429,7 @@ name_expr_with_subvar_value: STRING subv $$=N(POOL); O($$, OP_CREATE_EWPOOL); P($$, $1); - O($$, OP_WRITE); + O($$, OP_WRITE_VALUE); P($$, $2); O($$, OP_REDUCE_EWPOOL); }; @@ -322,16 +451,8 @@ subvar__get_write: '$' subvar_ref_name_r }; class_prefix: STRING ':' { - String& name=*SLA2S($1); - VClass *vclass=static_cast(PC->request->classes().get(name)); - if(!vclass) { - strcpy(PC->error, "'"); - strcat(PC->error, name.cstr()); - strcat(PC->error, "' class is undefined in call"); - YYERROR; - } - //TODO: убрать зависимость от статических @USE, сделать имя, а не ссылку - $$=CL(vclass); // vclass + $$=$1; // stack: class name string + O($$, OP_GET_CLASS); }; @@ -342,51 +463,74 @@ with: '$' name_without_curly_rdive '{' c O($$, OP_CREATE_RWPOOL); P($$, $4); O($$, OP_REDUCE_RWPOOL); - O($$, OP_WRITE); + O($$, OP_WRITE_VALUE); }; -/* expression */ +/* expr */ -expression_value: - empty_value /* optimized $var() case */ -| number /* optimized $var(STRING) case */ -| complex_expression /* $var(something complex) */ +expr_value: expr { + if(($$=$1)->size()==2) // only one string literal in there? + change_string_literal_to_double_literal($$); // make that string literal Double +}; +expr: + STRING +| get_value +| call_value +| '"' string_inside_quotes_value '"' { $$ = $2; } +| '(' expr ')' { $$ = $2; } +/* stack: operand // stack: @operand */ +| '-' expr %prec NEG { $$=$2; O($$, OP_NEG) } +| '~' expr { $$=$2; O($$, OP_INV) } +| '!' expr { $$=$2; O($$, OP_NOT) } +| "def" expr { $$=$2; O($$, OP_DEF) } +| "in" expr { $$=$2; O($$, OP_IN) } +| "-f" expr { $$=$2; O($$, OP_FEXISTS) } +/* stack: a,b // stack: a@b */ +| expr '-' expr { $$=$1; P($$, $3); O($$, OP_SUB) } +| expr '+' expr { $$=$1; P($$, $3); O($$, OP_ADD) } +| expr '*' expr { $$=$1; P($$, $3); O($$, OP_MUL) } +| expr '/' expr { $$=$1; P($$, $3); O($$, OP_DIV) } +| expr '%' expr { $$=$1; P($$, $3); O($$, OP_MOD) } +| expr '&' expr { $$=$1; P($$, $3); O($$, OP_BIN_AND) } +| expr '|' expr { $$=$1; P($$, $3); O($$, OP_BIN_OR) } +| expr '#' expr { $$=$1; P($$, $3); O($$, OP_BIN_XOR) } +| expr "&&" expr { $$=$1; P($$, $3); O($$, OP_LOG_AND) } +| expr "||" expr { $$=$1; P($$, $3); O($$, OP_LOG_OR) } +| expr "##" expr { $$=$1; P($$, $3); O($$, OP_LOG_XOR) } +| expr '<' expr { $$=$1; P($$, $3); O($$, OP_NUM_LT) } +| expr '>' expr { $$=$1; P($$, $3); O($$, OP_NUM_GT) } +| expr "<=" expr { $$=$1; P($$, $3); O($$, OP_NUM_LE) } +| expr ">=" expr { $$=$1; P($$, $3); O($$, OP_NUM_GE) } +| expr "==" expr { $$=$1; P($$, $3); O($$, OP_NUM_EQ) } +| expr "!=" expr { $$=$1; P($$, $3); O($$, OP_NUM_NE) } +| expr "lt" expr { $$=$1; P($$, $3); O($$, OP_STR_LT) } +| expr "gt" expr { $$=$1; P($$, $3); O($$, OP_STR_GT) } +| expr "le" expr { $$=$1; P($$, $3); O($$, OP_STR_LE) } +| expr "ge" expr { $$=$1; P($$, $3); O($$, OP_STR_GE) } +| expr "eq" expr { $$=$1; P($$, $3); O($$, OP_STR_EQ) } +| expr "ne" expr { $$=$1; P($$, $3); O($$, OP_STR_NE) } +| expr "is" expr { $$=$1; P($$, $3); O($$, OP_IS) } ; -complex_expression: expression_operand '*' expression_operand { - $$=$1; // stack: first operand - P($$, $3); // stack: first,second operands - O($$, OP_MUL); // value=first*second; stack: value -}; -expression_operand: number; -/* -complex_expression_value: complex_expression { - $$=N(POOL); - O($$, OP_CREATE_SWPOOL); /* stack: empty write context * / - P($$, $1); /* some codes to that context * / - O($$, OP_REDUCE_SWPOOL); /* context=pop; stack: context.get_string() * / +string_inside_quotes_value: maybe_codes { + $$=N(POOL); + O($$, OP_CREATE_SWPOOL); /* stack: empty write context */ + P($$, $1); /* some code that writes to that context */ + O($$, OP_REDUCE_SWPOOL); /* context=pop; stack: context.get_string() */ }; -*/ /* basics */ -write_str_literal: STRING { - if(SLA2S($1)->size()) { - $$=$1; - O($$, OP_WRITE); - } else { - // optimized case of special end of macro. see yylex - $$=N(POOL); - } -}; -number: STRING { - change_string_literal_to_double_literal($$=$1); +write_string: STRING { + // optimized from OP_STRING+OP_WRITE_VALUE to OP_STRING__WRITE + change_string_literal_to_write_string_literal($$=$1) }; -empty_value: /* empty */ { $$=SL(NEW VString(POOL)) }; +empty_string_value: /* empty */ { $$=VL(NEW VString(POOL)) }; empty: /* empty */ { $$=N(POOL) }; %% +#endif /* 000$111(2222)00 @@ -404,51 +548,89 @@ empty: /* empty */ { $$=N(POOL) }; 4:[^({]=pop */ -int yylex(YYSTYPE *lvalp, void *pc) { - #define lexical_brackets_nestage PC->brackets_nestages[PC->sp] +static int yylex(YYSTYPE *lvalp, void *pc) { + #define lexical_brackets_nestage PC.brackets_nestages[PC.sp] #define RC {result=c; goto break2; } register int c; int result; - if(PC->pending_state) { - result=PC->pending_state; - PC->pending_state=0; + if(PC.pending_state) { + result=PC.pending_state; + PC.pending_state=0; return result; } - char *begin=PC->source; - char *end; - int begin_line=PC->line; + const char *begin=PC.source; + const char *end; + int begin_line=PC.line; + int skip_analized=0; while(true) { - c=*(end=(PC->source++)); + c=*(end=(PC.source++)); if(c=='\n') { - PC->line++; - PC->col=0; + PC.line++; + PC.col=0; } else - PC->col++; + PC.col++; - // escaping: ^^ ^$ ^; ^) ^} ^( ^{ ^" - if(c=='^') - switch(*PC->source) { + if(c=='^' && PC.ls!=LS_COMMENT && PC.ls!=LS_DEF_COMMENT) + switch(*PC.source) { + // escaping: ^^ ^$ ^; ^) ^} ^( ^{ ^" case '^': case '$': case ';': case '[': case ']': case '{': case '}': - case '"': + case '"': if(end!=begin) { // append piece till ^ - PC->string->APPEND(begin, end-begin, PC->file, begin_line); + PC.string->APPEND(begin, end-begin, PC.file, begin_line); } - // reset piece 'start' position & line - begin=PC->source; // ^ - begin_line=PC->line; + // reset piece 'begin' position & line + begin=PC.source; // ^ + begin_line=PC.line; // skip over ^ and _ - PC->source++; PC->col++; + PC.source++; PC.col++; // skip analysis = forced literal continue; + + // converting ^#HH into char(hex(HH)) + case '#': + if(end!=begin) { + // append piece till ^ + PC.string->APPEND(begin, end-begin, PC.file, begin_line); + } + // #HH ? + if(PC.source[0]=='#' && PC.source[1] && PC.source[2]) { + char *hex=(char *)POOL.malloc(1); + hex[0]= + hex_value[(unsigned char)PC.source[1]]*0x10+ + hex_value[(unsigned char)PC.source[2]]; + if(hex[0]==0) { + result=BAD_HEX_LITERAL; + goto break2; // wrong hex value[no ^#00 chars allowed]: bail out + } + // append char(hex(HH)) + PC.string->APPEND(hex, 1, PC.file, begin_line); + // skip over ^#HH + PC.source+=3; + PC.col+=3; + // reset piece 'begin' position & line + begin=PC.source; // ^ + begin_line=PC.line; + continue; + } + break; + } + // #comment start skipping + if(c=='#' && PC.col==1) { + if(end!=begin) { + // append piece till # + PC.string->APPEND(begin, end-begin, PC.file, begin_line); } - switch(PC->ls) { + // fall into COMMENT lexical state [wait for \n] + push_LS(PC, LS_COMMENT); + } + switch(PC.ls) { // USER'S = NOT OURS case LS_USER: @@ -460,7 +642,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { push_LS(PC, LS_METHOD_NAME); RC; case '@': - if(PC->col==0+1) { + if(PC.col==0+1) { push_LS(PC, LS_DEF_NAME); RC; } @@ -468,6 +650,18 @@ int yylex(YYSTYPE *lvalp, void *pc) { } break; + // #COMMENT + case LS_COMMENT: + if(c=='\n') { + // skip comment + begin=PC.source; + begin_line=PC.line; + + pop_LS(PC); + continue; + } + break; + // STRING IN EXPRESSION case LS_EXPRESSION_STRING: switch(c) { @@ -487,10 +681,10 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_DEF_NAME: switch(c) { case '[': - PC->ls=LS_DEF_PARAMS; + PC.ls=LS_DEF_PARAMS; RC; case '\n': - PC->ls=LS_DEF_SPECIAL_BODY; + PC.ls=LS_DEF_SPECIAL_BODY; RC; } break; @@ -500,7 +694,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { case ';': RC; case ']': - PC->ls=*PC->source=='['?LS_DEF_LOCALS:LS_DEF_COMMENT; + PC.ls=*PC.source=='['?LS_DEF_LOCALS:LS_DEF_COMMENT; RC; case '\n': // wrong. bailing out pop_LS(PC); @@ -514,7 +708,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { case ';': RC; case ']': - PC->ls=LS_DEF_COMMENT; + PC.ls=LS_DEF_COMMENT; RC; case '\n': // wrong. bailing out pop_LS(PC); @@ -531,7 +725,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_DEF_SPECIAL_BODY: if(c=='\n') { - switch(*PC->source) { + switch(*PC.source) { case '@': case 0: // end of special_code pop_LS(PC); break; @@ -541,14 +735,18 @@ int yylex(YYSTYPE *lvalp, void *pc) { break; // (EXPRESSION) - case LS_EXPRESSION_BODY: + case LS_VAR_ROUND: + case LS_METHOD_ROUND: switch(c) { case ')': if(--lexical_brackets_nestage==0) - pop_LS(PC); //(EXPRESSION). + if(PC.ls==LS_METHOD_ROUND) // method round param ended + PC.ls=LS_METHOD_AFTER; // look for method end + else // PC.ls==LS_VAR_ROUND // variable constructor ended + pop_LS(PC); // return to normal life RC; case '$': - push_LS(PC, LS_VAR_NAME_IN_EXPRESSION); + push_LS(PC, LS_EXPRESSION_VAR_NAME); RC; case '^': push_LS(PC, LS_METHOD_NAME); @@ -556,49 +754,105 @@ int yylex(YYSTYPE *lvalp, void *pc) { case '(': lexical_brackets_nestage++; RC; - case '+': case '-': case '*': case '/': case '%': - case '&': case '|': - case '<': case '>': case '=': case '!': + case '-': + if(*PC.source=='f') { // -f + skip_analized=1; + result=FEXISTS; + } else + result=c; + goto break2; + case '+': case '*': case '/': case '%': + case '~': case ';': RC; + case '&': case '|': case '#': + if(*PC.source==c) { // && || + result=c=='#'?LXOR:c=='&'?LAND:LOR; + skip_analized=1; + } else + result=c; + goto break2; + case '<': case '>': case '=': case '!': + if(*PC.source=='=') { // <= >= == != + skip_analized=1; + switch(c) { + case '<': result=NLE; break; + case '>': result=NGE; break; + case '=': result=NEQ; break; + case '!': result=NNE; break; + } + } else + result=c; + goto break2; case '"': push_LS(PC, LS_EXPRESSION_STRING); RC; case 'l': case 'g': case 'e': case 'n': if(end==begin) // right after whitespace - switch(char next_c=*PC->source) { -// case '?': // ok [and bad cases, yacc would bark at them] - case 't': // lt gt [et nt] - case 'e': // le ge ne [ee] - case 'q': // eq [lq gq nq] - PC->source++; PC->col++; - PC->pending_state=next_c; - return c; + if(isspace(PC.source[1])) { + switch(*PC.source) { + // case '?': // ok [and bad cases, yacc would bark at them] + case 't': // lt gt [et nt] + result=c=='l'?SLT:c=='g'?SGT:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + case 'e': // le ge ne [ee] + result=c=='l'?SLE:c=='g'?SGE:c=='n'?SNE:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + case 'q': // eq [lq gq nq] + result=c=='e'?SEQ:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + } + } + break; + case 'i': + if(end==begin) // right after whitespace + if(isspace(PC.source[1])) { + switch(PC.source[0]) { + case 'n': // in + skip_analized=1; + result=IN; + goto break2; + case 's': // is + skip_analized=1; + result=IS; + goto break2; + } + } + break; + case 'd': + if(end==begin) // right after whitespace + if(PC.source[0]=='e' && PC.source[1]=='f') { // def + skip_analized=2; + result=DEF; + goto break2; } break; case ' ': case '\t': case '\n': - if(end!=begin) { - // append piece till whitespace - PC->string->APPEND(begin, end-begin, PC->file, begin_line); + if(end!=begin) { // there were a string after previous operator? + result=0; // return that string + goto break2; } - // reset piece 'start' position & line - begin=PC->source; // after whitespace - begin_line=PC->line; + // that's a leading|traling space or after-operator-space + // ignoring it + // reset piece 'begin' position & line + begin=PC.source; // after whitespace char + begin_line=PC.line; continue; } break; // VARIABLE GET/PUT/WITH case LS_VAR_NAME_SIMPLE: - case LS_VAR_NAME_IN_EXPRESSION: - if(PC->ls==LS_VAR_NAME_IN_EXPRESSION) { - // name in expression ends also before binary operators + case LS_EXPRESSION_VAR_NAME: + if(PC.ls==LS_EXPRESSION_VAR_NAME) { + // name in expr ends also before binary operators switch(c) { - case '+': case '-': case '*': case '/': case '%': - case '&': case '|': - case '<': case '>': case '=': case '!': + case '-': pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; } @@ -607,26 +861,32 @@ int yylex(YYSTYPE *lvalp, void *pc) { case 0: case ' ': case '\t': case '\n': case ';': - case ']': case '}': + case ']': case '}': case ')': case '"': + case '<': case '>': // these stand for HTML brackets and expression binary ops + case '+': case '*': case '/': case '%': + case '&': case '|': + case '=': case '!': + // common delimiters + case '\'': case ',': pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; case '[': - PC->ls=LS_VAR_SQUARE; + PC.ls=LS_VAR_SQUARE; lexical_brackets_nestage=1; RC; case '{': if(begin==end) { // ${name}, no need of EON, switching LS - PC->ls=LS_VAR_NAME_CURLY; + PC.ls=LS_VAR_NAME_CURLY; } else { - PC->ls=LS_VAR_CURLY; + PC.ls=LS_VAR_CURLY; lexical_brackets_nestage=1; } - + RC; case '(': - PC->ls=LS_EXPRESSION_BODY; + PC.ls=LS_VAR_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim @@ -694,11 +954,15 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_METHOD_NAME: switch(c) { case '[': - PC->ls=LS_METHOD_SQUARE; + PC.ls=LS_METHOD_SQUARE; lexical_brackets_nestage=1; RC; case '{': - PC->ls=LS_METHOD_CURLY; + PC.ls=LS_METHOD_CURLY; + lexical_brackets_nestage=1; + RC; + case '(': + PC.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim @@ -720,7 +984,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { RC; case ']': if(--lexical_brackets_nestage==0) { - PC->ls=LS_METHOD_AFTER; + PC.ls=LS_METHOD_AFTER; RC; } break; @@ -738,9 +1002,11 @@ int yylex(YYSTYPE *lvalp, void *pc) { case '^': push_LS(PC, LS_METHOD_NAME); RC; + case ';': // param delim + RC; case '}': if(--lexical_brackets_nestage==0) { - PC->ls=LS_METHOD_AFTER; + PC.ls=LS_METHOD_AFTER; RC; } break; @@ -751,18 +1017,23 @@ int yylex(YYSTYPE *lvalp, void *pc) { break; case LS_METHOD_AFTER: - if(c=='[') {/* )( }( */ - PC->ls=LS_METHOD_SQUARE; + if(c=='[') {/* ][ }[ )[ */ + PC.ls=LS_METHOD_SQUARE; lexical_brackets_nestage=1; RC; } - if(c=='{') {/* ){ }{ */ - PC->ls=LS_METHOD_CURLY; + if(c=='{') {/* ]{ }{ ){ */ + PC.ls=LS_METHOD_CURLY; + lexical_brackets_nestage=1; + RC; + } + if(c=='(') {/* ]( }( )( */ + PC.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; } pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; } @@ -773,41 +1044,36 @@ int yylex(YYSTYPE *lvalp, void *pc) { } break2: - if(begin==end) - return result; - else { - PC->pending_state=result; - // strip last \n before LS_DEF_NAME or EOF - if((c=='@' || c==0) && end[-1]=='\n') + if(end!=begin) { // there is last piece? + if((c=='@' || c==0) && end[-1]=='\n') { // we are before LS_DEF_NAME or EOF? + // strip last \n end--; - if(end!=begin) { - // append last piece - PC->string->APPEND(begin, end-begin, PC->file, begin_line/*, start_col*/); } + if(end!=begin) { // last piece still alive? + // append it + PC.string->APPEND(begin, end-begin, PC.file, begin_line/*, start_col*/); + } + } + if(PC.string->size()) { // something accumulated? // create STRING value: array of OP_VALUE+vstring - *lvalp=SL(NEW VString(*PC->string)); + *lvalp=VL(NEW VString(*PC.string)); // new pieces storage - PC->string=NEW String(POOL); - // go! - return STRING; + PC.string=NEW String(POOL); + // make current result be pending for next call, return STRING for now + PC.pending_state=result; result=STRING; } + if(skip_analized) { + PC.source+=skip_analized; PC.col+=skip_analized; + } + return result; } -int real_yyerror(parse_control *pc, char *s) /* Called by yyparse on error */ - { - //fprintf(stderr, "[%s]\n", s); - - strncpy(pc->error, s, MAX_STRING); // TODO: перепроверить с треклятым последним байтом +static int real_yyerror(parse_control *pc, char *s) { // Called by yyparse on error + strncpy(PC.error, s, MAX_STRING); return 1; - } - -static void - yyprint( - FILE *file, - int type, - YYSTYPE value) - { - if(type==STRING) - fprintf(file, " \"%s\"", SLA2S(value)->cstr()); - } +} +static void yyprint(FILE *file, int type, YYSTYPE value) { + if(type==STRING) + fprintf(file, " \"%s\"", SLA2S(value)->cstr()); +}