--- parser3/src/main/compile.y 2003/03/27 14:51:28 1.204.2.8.2.7 +++ parser3/src/main/compile.y 2003/04/03 06:16:05 1.204.2.8.2.10 @@ -5,7 +5,7 @@ Copyright (c) 2001, 2003 ArtLebedev Group (http://www.artlebedev.com) Author: Alexander Petrosyan (http://design.ru/paf) - $Id: compile.y,v 1.204.2.8.2.7 2003/03/27 14:51:28 paf Exp $ + $Id: compile.y,v 1.204.2.8.2.10 2003/04/03 06:16:05 paf Exp $ */ /** @@ -24,9 +24,11 @@ #define YYLEX_PARAM pc #define YYDEBUG 1 #define YYERROR_VERBOSE 1 -#define yyerror(msg) real_yyerror((parse_control *)pc, msg) +#define yyerror(msg) real_yyerror((Parse_control *)pc, msg) #define YYPRINT(file, type, value) yyprint(file, type, value) +// includes + #include "compile_tools.h" #include "pa_value.h" #include "pa_request.h" @@ -36,16 +38,21 @@ #include "pa_vvoid.h" #include "pa_vmethod_frame.h" +// defines + #define USE_CONTROL_METHOD_NAME "USE" +#define TAB_SIZE 8 + +// forwards -static int real_yyerror(parse_control *pc, char *s); +static int real_yyerror(Parse_control *pc, char *s); static void yyprint(FILE *file, int type, YYSTYPE value); static int yylex(YYSTYPE *lvalp, void *pc); // local convinient inplace typecast & var #undef PC -#define PC (*(parse_control *)pc) +#define PC (*(Parse_control *)pc) #undef POOL #define POOL (*PC.pool) #ifndef DOXYGEN @@ -498,7 +505,7 @@ class_prefix: ; class_static_prefix: STRING ':' { $$=$1; // stack: class name string - if(*LA2S(*$$) == BASE_NAME) { // pseude BASE class + if(*LA2S(*$$) == BASE_NAME) { // pseudo BASE class if(VStateless_class* base=PC.cclass->base_class()) { change_string_literal_value(*$$, base->name()); } else { @@ -585,7 +592,7 @@ write_string: STRING { change_string_literal_to_write_string_literal(*($$=$1)) }; -void_value: /* empty */ { $$=VL(new VVoid()) }; +void_value: /* empty */ { $$=VL(new VVoid(), 0, 0, 0) }; empty: /* empty */ { $$=N() }; %% @@ -607,55 +614,67 @@ empty: /* empty */ { $$=N() }; 4:[^({]=pop */ -static int yylex(YYSTYPE *lvalp, void *pc) { - #define lexical_brackets_nestage PC.brackets_nestages[PC.ls_sp] +inline ungetc(Parse_control& pc, uint last_line_end_col) { + pc.source--; + if(pc.pos.col==0) { + --pc.pos.line; pc.pos.col=last_line_end_col; + } else + --pc.pos.col; + +} +static int yylex(YYSTYPE *lvalp, void *apc) { + register Parse_control& pc=*static_cast(apc); + + #define lexical_brackets_nestage pc.brackets_nestages[pc.ls_sp] #define RC {result=c; goto break2; } - register int c; - int result; + 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; } - const char *begin=PC.source; + const char *begin=pc.source; + Pos begin_pos=pc.pos; + uint last_line_end_col=0; const char *end; - int begin_line=PC.line; int skip_analized=0; while(true) { - c=*(end=(PC.source++)); -// fprintf(stderr, "\nchar: %c %02X; nestage: %d, sp=%d", c, c, lexical_brackets_nestage, PC.sp); + c=*(end=(pc.source++)); +// fprintf(stderr, "\nchar: %c %02X; nestage: %d, sp=%d", c, c, lexical_brackets_nestage, pc.sp); if(c=='\n') { - PC.line++; - PC.col=0; + pc.pos.line++; + last_line_end_col=pc.pos.col; + pc.pos.col=0; } else - PC.col++; + pc.pos.col+=c=='\t'?TAB_SIZE:1; - if(c=='@' && PC.col==0+1) { - if(PC.ls==LS_DEF_SPECIAL_BODY) { + if(c=='@' && pc.pos.col==0+1) { + if(pc.ls==LS_DEF_SPECIAL_BODY) { // @SPECIAL // ... // @punctuation - begin_line=PC.line; + begin=pc.source; // ->punctuation + begin_pos=pc.pos; // skip over _ after ^ - PC.source++; PC.col++; + pc.source++; pc.pos.col++; // skip analysis = forced literal continue; // converting ^#HH into char(hex(HH)) case '#': if(end!=begin) { + if(!pc.string_start) + pc.string_start=begin_pos; // append piece till ^ - PC.string.append_strdup(begin, end-begin); + pc.string.append_strdup(begin, end-begin); } // #HH ? - if(PC.source[0]=='#' && PC.source[1] && PC.source[2]) { + if(pc.source[0]=='#' && pc.source[1] && pc.source[2]) { char c= - hex_value[(unsigned char)PC.source[1]]*0x10+ - hex_value[(unsigned char)PC.source[2]]; + hex_value[(unsigned char)pc.source[1]]*0x10+ + hex_value[(unsigned char)pc.source[2]]; if(c==0) { result=BAD_HEX_LITERAL; goto break2; // wrong hex value[no ^#00 chars allowed]: bail out } // append char(hex(HH)) - PC.string.append(c); + pc.string.append(c); // skip over ^#HH - PC.source+=3; - PC.col+=3; + pc.source+=3; + pc.pos.col+=3; // reset piece 'begin' position & line - begin=PC.source; // ->after ^#HH - begin_line=PC.line; + begin=pc.source; // ->after ^#HH + begin_pos=pc.pos; // skip analysis = forced literal continue; } @@ -718,45 +741,47 @@ default: } } // #comment start skipping - if(c=='#' && PC.col==1) { + if(c=='#' && pc.pos.col==1) { if(end!=begin) { + if(!pc.string_start) + pc.string_start=begin_pos; // append piece till # - PC.string.append_strdup(begin, end-begin); + pc.string.append_strdup(begin, end-begin); } // fall into COMMENT lexical state [wait for \n] - push_LS(PC, LS_USER_COMMENT); + push_LS(pc, LS_USER_COMMENT); continue; } - switch(PC.ls) { + switch(pc.ls) { // USER'S = NOT OURS case LS_USER: - case LS_NAME_SQUARE_PART: // name.[here].xxx - if(PC.trim_bof) + case LS_NAME_SQUARE_PART: // name.[here].xxx + if(pc.trim_bof) switch(c) { case '\n': case ' ': case '\t': - begin=PC.source; - begin_line=PC.line; + begin=pc.source; + begin_pos=pc.pos; continue; // skip it default: - PC.trim_bof=false; + pc.trim_bof=false; } switch(c) { case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; case ']': - if(PC.ls==LS_NAME_SQUARE_PART) + if(pc.ls==LS_NAME_SQUARE_PART) if(--lexical_brackets_nestage==0) {// $name.[co<]?>de<]?> - pop_LS(PC); // $name.[co<]>de<]!> + pop_LS(pc); // $name.[co<]>de<]!> RC; } break; case '[': // $name.[co<[>de] - if(PC.ls==LS_NAME_SQUARE_PART) + if(pc.ls==LS_NAME_SQUARE_PART) lexical_brackets_nestage++; break; } @@ -766,10 +791,10 @@ default: case LS_USER_COMMENT: if(c=='\n') { // skip comment - begin=PC.source; - begin_line=PC.line; + begin=pc.source; + begin_pos=pc.pos; - pop_LS(PC); + pop_LS(pc); continue; } break; @@ -781,17 +806,17 @@ default: case '"': case '\'': if( - PC.ls == LS_EXPRESSION_STRING_QUOTED && c=='"' || - PC.ls == LS_EXPRESSION_STRING_APOSTROFED && c=='\'') { - pop_LS(PC); //"abc". | 'abc'. + pc.ls == LS_EXPRESSION_STRING_QUOTED && c=='"' || + pc.ls == LS_EXPRESSION_STRING_APOSTROFED && c=='\'') { + pop_LS(pc); //"abc". | 'abc'. RC; } break; case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; } break; @@ -800,10 +825,10 @@ default: 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; @@ -816,10 +841,10 @@ default: 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); + pop_LS(pc); RC; } break; @@ -830,17 +855,17 @@ default: case ';': RC; case ']': - PC.ls=LS_DEF_COMMENT; + pc.ls=LS_DEF_COMMENT; RC; case '\n': // wrong. bailing out - pop_LS(PC); + pop_LS(pc); RC; } break; case LS_DEF_COMMENT: if(c=='\n') { - pop_LS(PC); + pop_LS(pc); RC; } break; @@ -856,31 +881,33 @@ default: switch(c) { case ')': if(--lexical_brackets_nestage==0) - 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 + 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 '#': // comment start skipping if(end!=begin) { + if(!pc.string_start) + pc.string_start=begin_pos; // append piece till # - PC.string.append_strdup(begin, end-begin); + pc.string.append_strdup(begin, end-begin); } // fall into COMMENT lexical state [wait for \n] - push_LS(PC, LS_EXPRESSION_COMMENT); + push_LS(pc, LS_EXPRESSION_COMMENT); lexical_brackets_nestage=1; continue; case '$': - push_LS(PC, LS_EXPRESSION_VAR_NAME_WITH_COLON); + push_LS(pc, LS_EXPRESSION_VAR_NAME_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; case '(': lexical_brackets_nestage++; RC; case '-': - switch(*PC.source) { + switch(*pc.source) { case 'f': // -f skip_analized=1; result=FEXISTS; @@ -899,17 +926,17 @@ default: case ';': RC; case '&': case '|': - if(*PC.source==c) { // && || + if(*pc.source==c) { // && || result=c=='&'?LAND:LOR; skip_analized=1; } else result=c; goto break2; case '!': - switch(PC.source[0]) { + switch(pc.source[0]) { case '|': // !| !|| skip_analized=1; - if(PC.source[1]=='|') { + if(pc.source[1]=='|') { skip_analized++; result=LXOR; } else @@ -923,7 +950,7 @@ default: RC; case '<': // <<, <=, < - switch(*PC.source) { + switch(*pc.source) { case '<': // <[<] skip_analized=1; result=NSL; break; case '=': // <[=] @@ -933,7 +960,7 @@ default: } goto break2; case '>': // >>, >=, > - switch(*PC.source) { + switch(*pc.source) { case '>': // >[>] skip_analized=1; result=NSR; break; case '=': // >[=] @@ -943,7 +970,7 @@ default: } goto break2; case '=': // == - switch(*PC.source) { + switch(*pc.source) { case '=': // =[=] skip_analized=1; result=NEQ; break; default: // =[] @@ -952,15 +979,15 @@ default: goto break2; case '"': - push_LS(PC, LS_EXPRESSION_STRING_QUOTED); + push_LS(pc, LS_EXPRESSION_STRING_QUOTED); RC; case '\'': - push_LS(PC, LS_EXPRESSION_STRING_APOSTROFED); + push_LS(pc, LS_EXPRESSION_STRING_APOSTROFED); RC; case 'l': case 'g': case 'e': case 'n': if(end==begin) // right after whitespace - if(isspace(PC.source[1])) { - switch(*PC.source) { + 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; @@ -979,8 +1006,8 @@ default: break; case 'i': if(end==begin) // right after whitespace - if(isspace(PC.source[1])) { - switch(PC.source[0]) { + if(isspace(pc.source[1])) { + switch(pc.source[0]) { case 'n': // in skip_analized=1; result=IN; @@ -994,7 +1021,7 @@ default: break; case 'd': if(end==begin) // right after whitespace - if(PC.source[0]=='e' && PC.source[1]=='f') { // def + if(pc.source[0]=='e' && pc.source[1]=='f') { // def skip_analized=2; result=DEF; goto break2; @@ -1008,8 +1035,8 @@ default: // 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; + begin=pc.source; // after whitespace char + begin_pos=pc.pos; continue; } break; @@ -1017,17 +1044,17 @@ default: if(c=='(') lexical_brackets_nestage++; - switch(*PC.source) { + switch(*pc.source) { case '\n': case ')': - if(*PC.source==')') + if(*pc.source==')') if(--lexical_brackets_nestage!=0) continue; // skip comment - begin=PC.source; - begin_line=PC.line; + begin=pc.source; + begin_pos=pc.pos; - pop_LS(PC); + pop_LS(pc); continue; } break; @@ -1038,28 +1065,28 @@ default: case LS_EXPRESSION_VAR_NAME_WITH_COLON: case LS_EXPRESSION_VAR_NAME_WITHOUT_COLON: if( - PC.ls==LS_EXPRESSION_VAR_NAME_WITH_COLON || - PC.ls==LS_EXPRESSION_VAR_NAME_WITHOUT_COLON) { + pc.ls==LS_EXPRESSION_VAR_NAME_WITH_COLON || + pc.ls==LS_EXPRESSION_VAR_NAME_WITHOUT_COLON) { // name in expr ends also before switch(c) { // expression minus case '-': // expression integer division case '\\': - pop_LS(PC); - PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } + pop_LS(pc); + ungetc(pc, last_line_end_col); result=EON; goto break2; } } if( - PC.ls==LS_VAR_NAME_SIMPLE_WITHOUT_COLON || - PC.ls==LS_EXPRESSION_VAR_NAME_WITHOUT_COLON) { + pc.ls==LS_VAR_NAME_SIMPLE_WITHOUT_COLON || + pc.ls==LS_EXPRESSION_VAR_NAME_WITHOUT_COLON) { // name already has ':', stop before next switch(c) { case ':': - pop_LS(PC); - PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } + pop_LS(pc); + ungetc(pc, last_line_end_col); result=EON; goto break2; } @@ -1078,45 +1105,45 @@ default: case ',': case '?': case '#': // before call case '^': - pop_LS(PC); - PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } + pop_LS(pc); + ungetc(pc, last_line_end_col); result=EON; goto break2; case '[': // $name.<[>code] - if(PC.col>1/*not first column*/ && ( + if(pc.pos.col>1/*not first column*/ && ( end[-1]=='$'/*was start of get*/ || end[-1]==':'/*was class name delim */ || end[-1]=='.'/*was name delim */ )) { - push_LS(PC, LS_NAME_SQUARE_PART); + push_LS(pc, LS_NAME_SQUARE_PART); lexical_brackets_nestage=1; RC; } - PC.ls=LS_VAR_SQUARE; + pc.ls=LS_VAR_SQUARE; lexical_brackets_nestage=1; RC; case '{': if(begin==end) { // ${name}, no need of EON, switching LS - PC.ls=LS_VAR_NAME_CURLY; + pc.ls=LS_VAR_NAME_CURLY; } else { - PC.ls=LS_VAR_CURLY; + pc.ls=LS_VAR_CURLY; lexical_brackets_nestage=1; } RC; case '(': - PC.ls=LS_VAR_ROUND; + pc.ls=LS_VAR_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim case '$': // name part subvar case ':': // class<:>name // go to _WITHOUT_COLON state variant... - if(PC.ls==LS_VAR_NAME_SIMPLE_WITH_COLON) - PC.ls=LS_VAR_NAME_SIMPLE_WITHOUT_COLON; - else if(PC.ls==LS_EXPRESSION_VAR_NAME_WITH_COLON) - PC.ls=LS_EXPRESSION_VAR_NAME_WITHOUT_COLON; + if(pc.ls==LS_VAR_NAME_SIMPLE_WITH_COLON) + pc.ls=LS_VAR_NAME_SIMPLE_WITHOUT_COLON; + else if(pc.ls==LS_EXPRESSION_VAR_NAME_WITH_COLON) + pc.ls=LS_EXPRESSION_VAR_NAME_WITHOUT_COLON; // ...stop before next ':' RC; } @@ -1126,11 +1153,11 @@ default: switch(c) { case '[': // ${name.<[>code]} - push_LS(PC, LS_NAME_SQUARE_PART); + push_LS(pc, LS_NAME_SQUARE_PART); lexical_brackets_nestage=1; RC; case '}': // ${name} finished, restoring LS - pop_LS(PC); + pop_LS(pc); RC; case '.': // name part delim case '$': // name part subvar @@ -1142,14 +1169,14 @@ default: case LS_VAR_SQUARE: switch(c) { case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; case ']': if(--lexical_brackets_nestage==0) { - pop_LS(PC); + pop_LS(pc); RC; } break; @@ -1164,14 +1191,14 @@ default: case LS_VAR_CURLY: switch(c) { case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; case '}': if(--lexical_brackets_nestage==0) { - pop_LS(PC); + pop_LS(pc); RC; } break; @@ -1186,24 +1213,24 @@ default: switch(c) { case '[': // ^name.<[>code].xxx - if(PC.col>1/*not first column*/ && ( + if(pc.pos.col>1/*not first column*/ && ( end[-1]=='^'/*was start of call*/ || // never, ^[ is literal... end[-1]==':'/*was class name delim */ || end[-1]=='.'/*was name delim */ )) { - push_LS(PC, LS_NAME_SQUARE_PART); + push_LS(pc, LS_NAME_SQUARE_PART); lexical_brackets_nestage=1; RC; } - 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; + pc.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim @@ -1218,16 +1245,16 @@ default: case LS_METHOD_SQUARE: switch(c) { case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + push_LS(pc, LS_METHOD_NAME); RC; case ';': // param delim RC; case ']': if(--lexical_brackets_nestage==0) { - PC.ls=LS_METHOD_AFTER; + pc.ls=LS_METHOD_AFTER; RC; } break; @@ -1240,16 +1267,16 @@ default: case LS_METHOD_CURLY: switch(c) { case '$': - push_LS(PC, LS_VAR_NAME_SIMPLE_WITH_COLON); + push_LS(pc, LS_VAR_NAME_SIMPLE_WITH_COLON); RC; case '^': - push_LS(PC, LS_METHOD_NAME); + 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; @@ -1261,22 +1288,22 @@ default: case LS_METHOD_AFTER: if(c=='[') {/* ][ }[ )[ */ - PC.ls=LS_METHOD_SQUARE; + pc.ls=LS_METHOD_SQUARE; lexical_brackets_nestage=1; RC; } if(c=='{') {/* ]{ }{ ){ */ - PC.ls=LS_METHOD_CURLY; + pc.ls=LS_METHOD_CURLY; lexical_brackets_nestage=1; RC; } if(c=='(') {/* ]( }( )( */ - PC.ls=LS_METHOD_ROUND; + pc.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; } - pop_LS(PC); - PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } + pop_LS(pc); + ungetc(pc, last_line_end_col); result=EON; goto break2; } @@ -1294,26 +1321,31 @@ break2: if(end!=begin && end[-1]=='\n') // allow one empty line before LS_DEF_NAME end--; } - if(end!=begin && PC.ls!=LS_USER_COMMENT) { // last piece still alive and not comment? + if(end!=begin && pc.ls!=LS_USER_COMMENT) { // last piece still alive and not comment? + if(!pc.string_start) + pc.string_start=begin_pos; // append it - PC.string.append_strdup(begin, end-begin); + pc.string.append_strdup(begin, end-begin); } } - if(!PC.string.is_empty()) { // something accumulated? + if(!pc.string.is_empty()) { // something accumulated? // create STRING value: array of OP_VALUE+vstring - *lvalp=VL(new VString(*new String(PC.string, String::L_CLEAN))); + *lvalp=VL( + new VString(*new String(pc.string, String::L_CLEAN)), + pc.file_no, pc.string_start.line, pc.string_start.col); // new pieces storage - PC.string.clear(); + pc.string.clear(); + pc.string_start.clear(); // make current result be pending for next call, return STRING for now - PC.pending_state=result; result=STRING; + pc.pending_state=result; result=STRING; } if(skip_analized) { - PC.source+=skip_analized; PC.col+=skip_analized; + pc.source+=skip_analized; pc.pos.col+=skip_analized; } return result; } -static int real_yyerror(parse_control *pc, char *s) { // Called by yyparse on error +static int real_yyerror(Parse_control *pc, char *s) { // Called by yyparse on error strncpy(PC.error, s, MAX_STRING); return 1; }