Index: tools/prettyprinter/parse.cc =================================================================== --- tools/prettyprinter/parse.cc (revision 7d4f6ed5e24a81ba67229ed1240413b39513e433) +++ (revision ) @@ -1,94 +1,0 @@ -// -*- Mode: C++ -*- -// -// Pretty Printer, Copyright (C) Richard C. Bilson and Rodolfo G. Esteves 2001 -// -// parse.cc -- -// -// Author : Richard C. Bilson and Rodolfo G. Esteves -// Created On : Sun Dec 16 15:06:31 2001 -// Last Modified By : Peter A. Buhr -// Last Modified On : Tue Apr 9 22:38:19 2002 -// Update Count : 435 -// - -#include "parse.h" -#include "filter.h" -#include "yacc.tab.h" - - -Token::Token( const string &text, int kind ) : text(text), kind(kind) { - left = down = NULL; -} // Token::Token - -Token::Token( const string &text, list &ws_list, int kind ) : text(text), kind(kind) { -// cerr << "Token3 : text \"" << text << "\""; -// for ( list::iterator i = ws_list.begin(); i != ws_list.end(); i ++ ) { -// cerr << " WSt: \"" << *i << "\""; -// } -// cerr << endl; - left = down = NULL; - // O(1) transfer of the current lex whitespace list to the token's - // whitespace list, and clearing the lex whitespace list. - Token::ws_list.splice( Token::ws_list.end(), ws_list ); -} // Token::Token - -void Token::addLeftTail( Token *n ) { - Token *p = this; - while ( p->left != 0 ) { - p = p->left; - } // while - p->left = n; -} // Token::addLeftTail - -void Token::addDownLeftTail( Token *n ) { - if ( down == 0 ) { - down = n; - } else { - down->addLeftTail( n ); - } // if -} // Token::addDownLeftTail - -bool Token::isTerminal() { - return kind < END_TERMINALS; -} // Token::isTerminal - -int Token::getKind() const { - return kind; -} // Token::getKind() - -string Token::getText() const { - return text; -} // Token::getText() - -string Token::getWS() { - string ret; - // concatenate whitespace and comment text - for ( list::iterator i = ws_list.begin(); i != ws_list.end(); i ++ ) { - ret += *i; - } // for - return ret; -} // Token::getWS - -bool Token::isComment() { - return ws_list.size() > 1 || ( ws_list.size() == 1 && (*ws_list.begin())[0] == '/' ); -} // Token::isComment - -string Token::getComment() { - string ret; - // concatenate whitespace and comment text up to the last comment - if ( ws_list.size() > 1 ) { - list::iterator end = -- ws_list.end(); - if ( (*end)[0] == '/' ) end ++; - for ( list::iterator i = ws_list.begin(); i != end; i ++ ) { - ret += *i; - } // for - } else if ( ws_list.size() == 1 ) { - if ( (*ws_list.begin())[0] == '/' ) ret = *ws_list.begin(); - } // if - return ret; -} // Token::getComment - - -// Local Variables: // -// compile-command: "gmake" // -// End: // Index: tools/prettyprinter/parse.h =================================================================== --- tools/prettyprinter/parse.h (revision 7d4f6ed5e24a81ba67229ed1240413b39513e433) +++ (revision ) @@ -1,57 +1,0 @@ -// -*- Mode: C++ -*- -// -// Pretty Printer, Copyright (C) Richard C. Bilson and Rodolfo G. Esteves 2001 -// -// parse.h - Various declarations that are needed so that the generated parser -// and lexer compile with C++, and to share information between the -// parser, lexer, and driver program -// -// Author : Richard C. Bilson and Rodolfo G. Esteves -// Created On : Sun Dec 16 15:00:49 2001 -// Last Modified By : Peter A. Buhr -// Last Modified On : Tue Dec 17 10:55:31 2002 -// Update Count : 163 -// - -#ifndef __PARSE_H__ -#define __PARSE_H__ - -#define YYDEBUG_LEXER_TEXT( yylval ) // lexer loads this up each time -#define YYDEBUG 1 // get the pretty debugging code to compile - -int yylex(); - -#include -#include -#include - -using std::cout; -using std::cerr; -using std::endl; -using std::string; -using std::list; - -struct Token { - string text; // text of terminal or non-terminal token - int kind; // kind of terminal or non-terminal token - list ws_list; // list of whitespace and comments before terminal token - Token *left, *down; // binary parse tree links - - Token( const string &, int ); - Token( const string &, list &, int ); - void addLeftTail( Token * ); - void addDownLeftTail( Token * ); - bool isTerminal(); - int getKind() const; - string getText() const; - string getWS(); - bool isComment(); - string getComment(); -}; - -#endif /* __PARSE_H__ */ - - -// Local Variables: // -// compile-command: "gmake" // -// End: // Index: tools/prettyprinter/yacc.yy =================================================================== --- tools/prettyprinter/yacc.yy (revision 7d4f6ed5e24a81ba67229ed1240413b39513e433) +++ (revision ) @@ -1,481 +1,0 @@ -/* -*- Mode: C -*- - * - * Pretty Printer, Copyright (C) Rodolfo G. Esteves and Peter A. Buhr 2001 - * Permission is granted to copy this grammar and to use it within software systems. - * THIS GRAMMAR IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES. - * - * yacc.y -- - * - * Author : Rodolfo G. Esteves - * Created On : Sat Dec 15 13:44:21 2001 - * Last Modified By : Peter A. Buhr - * Last Modified On : Mon Jun 27 21:51:06 2016 - * Update Count : 1028 - */ - -%{ -#include -#include "parse.h" -#include "filter.h" - -#define YYDEBUG 1 // get the pretty debugging code to compile - -extern list ws_list; // lex variable containing accumulated whitespace -void lexC( void ); -string lexYacc( void ); - -void yyerror( char *s ) { - extern int yylineno; - - cerr << "Error in line: " << yylineno << ": " << s << endl; - return; -} - -Token *declstart; -Token *rulestart; -Token *nameliststart; -%} - -%union { - Token *tokenp; -} - -%token ',' -%token '<' -%token '>' -%token '{' -%token '}' -%token ':' -%token ';' -%token '|' - -%token MARK // %% -%token LCURL // %{ -%token RCURL // %} - -%token INTEGER // integer constant -%token CHARACTER // character constant -%token IDENTIFIER // identifier -%token CODE // C code - -%token START // %start -%token UNION // %union -%token TOKEN // %token -%token LEFT // %left -%token RIGHT // %right -%token NONASSOC // %nonassoc -%token TYPE // %type -%token PURE_PARSER // %pure_parser -%token SEMANTIC_PARSER // %semantic_parser -%token EXPECT // %expect -%token THONG // %thong - -%token PREC // %prec - -%token END_TERMINALS // ALL TERMINAL TOKEN NAMES MUST APPEAR BEFORE THIS - -%type sections -%token _SECTIONS -%type mark -%type defsection_opt -%token _DEFSECTION_OPT -%type declarations -%type literalblock -%token _LITERALBLOCK -%type declaration -%token _DECLARATION -%type union -%type rword -%type tag_opt -%token _TAG_OPT -%type namenolist -%token _NAMENOLIST -%type nameno -%token _NAMENO -%type namelist -%token _NAMELIST -%type name -%type rulesection -%token _RULESECTION -%type rules -%token _RULE -%type lhs -%token _LHS -%type rhs -%token _RHS -%type prod -%type prec -%token _PREC -%type action -%token _ACTION -%type usersection_opt -%token _USERSECTION_OPT -%type ccode_opt -%type blocks - -%start grammar - -%% -grammar : sections - { - filter( $1 ); /* filter parse tree */ - freeTree( $1 ); /* free parse-tree storage (optional: used with purify) */ - } - ; - -sections : defsection_opt mark rulesection usersection_opt - { - $$ = new Token( "sections", _SECTIONS ); - $1->left = $2; - $2->left = $3; - $3->left = $4; - $$->down = $1; - } - ; - -mark : MARK - | error /* missing %% */ - { - cerr << "no input grammar, missing %% mark" << endl; - exit( -1 ); - } - ; - -defsection_opt : /* empty */ - { - //cerr << "defsection_opt1: " << endl; - $$ = new Token( "declaration_opt", _DEFSECTION_OPT ); - } - | declarations - { - //cerr << "defsection_opt2: " << $1->text << "(" << $1 << ")" << endl; - $$ = new Token( "declaration_opt", _DEFSECTION_OPT ); - $$->down = declstart; - } - ; - -declarations : literalblock - { - //cerr << "declarations1: " << $1->text << "(" << $1 << ")" << endl; - $$ = declstart = $1; - } - | declaration - { - //cerr << "declarations2: " << $1->text << "(" << $1 << ")" << endl; - $$ = declstart = new Token( "declaration", _DECLARATION ); - $$->down = $1; - } - | declarations literalblock - { - //cerr << "declarations3: "<< $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->left = $2; - $$ = $2; - } - | declarations declaration - { - //cerr << "declarations4: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $$ = new Token( "declaration", _DECLARATION ); - $1->left = $$; - $$->down = $2; - } - ; - -literalblock : LCURL - { lexC(); } - ccode_opt - { $$ = new Token( lexYacc(), CODE ); } - RCURL - { - //cerr << "literalblock: " << $1->text << "(" << $1 << ") " << $4->text << " " << $5->text << "(" << $5 << ")" << endl; - $1->left = $4; - $4->left = $5; - $$ = new Token( "literalblock", _LITERALBLOCK ); - $$->down = $1; - } - ; - -declaration : union - | START IDENTIFIER - { - $1->left = $2; - $$ = $1; - } - | rword tag_opt namenolist - { - Token *n = new Token( "namenolist", _NAMENOLIST ); - n->down = nameliststart; - $1->left = $2; - $2->left = n; - $$ = $1; - } - | TYPE tag_opt namelist - { - Token *n = new Token( "namelist", _NAMELIST ); - n->down = nameliststart; - $1->left = $2; - $2->left = n; - $$ = $1; - } - | PURE_PARSER - | SEMANTIC_PARSER - | EXPECT INTEGER /* bison */ - { - $1->left = $2; - $$ = $1; - } - | THONG /* bison */ - ; - -union : UNION - '{' - { lexC(); } - ccode_opt - { - // Remove the trailing '}' which is added in lex. - string temp( lexYacc() ); - $$ = new Token( temp.substr( 0, temp.length() - 1 ), CODE ); - } - '}' - { - $1->left = $2; - $2->left = $5; - $5->left = $6; - $$ = $1; - } - ; - -rword : TOKEN - | LEFT - | RIGHT - | NONASSOC - ; - -tag_opt : /* empty */ - { - //cerr << "tag_opt" << endl; - $$ = new Token( "tag_opt", _TAG_OPT ); - } - | '<' IDENTIFIER '>' - { - $1->left = $2; - $2->left = $3; - $$ = new Token( "tag_opt", _TAG_OPT ); - $$->down = $1; - } - ; - -namenolist : nameno - { - //cerr << "namenolist1: " << $1->text << "(" << $1 << ")" << endl; - $$ = nameliststart = $1; - } - | namenolist nameno - { - //cerr << "namenolist2: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->left = $2; - $$ = $2; - } - | namenolist ',' nameno - { - //cerr << "namenolist3: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ") " << $3->text << "(" << $3 << ")" << endl; - $1->left = $2; - $2->left = $3; - $$ = $3; - } - ; - -nameno : name - { - $$ = new Token( "nameno", _NAMENO ); - $$->down = $1; - } - | name INTEGER - { - $$ = new Token( "nameno", _NAMENO ); - $1->left = $2; - $$->down = $1; - } - ; - -namelist : name - { - //cerr << "namelist1: " << $1->text << "(" << $1 << ")" << endl; - $$ = nameliststart = $1; - } - | namelist name - { - //cerr << "namelist2: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->left = $2; - $$ = $2; - } - | namelist ',' name - { - //cerr << "namelist3: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ") " << $3->text << "(" << $3 << ")" << endl; - $1->left = $2; - $2->left = $3; - $$ = $3; - } - ; - -name : IDENTIFIER - | CHARACTER - ; - -rulesection : rules - { - //cerr << "rulesection1: " << $1->text << "(" << $1 << ")" << endl; - $$ = new Token( "rulesection", _RULESECTION ); - $$->down = $1; - } - | error /* no rules */ - { - cerr << "no rules in the input grammar" << endl; - exit( -1 ); - } - ; - -// These grammar rules are complex because the Yacc language is LR(2) due to the optional ';' at the end of -// rules. The following rules convert the LR(2) grammar into LR(1) by lengthening the rules to allow -// sufficient look ahead. Unfortunately, this change makes handling the semantic actions more complex because -// there are two lists (rules, rhs) being built but only one list tail can be returned through $$ for -// chaining. - -rules : lhs rhs - { - //cerr << "rules1: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $$ = rulestart; - } - | lhs rhs ';' - { - //cerr << "rules2: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ") " << $3->text << "(" << $3 << ")" << endl; - $2->addDownLeftTail( $3 ); - $$ = rulestart; - } - ; - -lhs : IDENTIFIER ':' - { - //cerr << "lhs: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $$ = new Token( "lhs", _LHS ); - //cerr << " lhs: " << $$->text << "(" << $$ << ")" << endl; - $1->left = $2; - $$->down = $1; - } - ; - -rhs : /* empty */ - { - //cerr << "rhs1: " << $0->text << "(" << $0 << ")" << endl; - rulestart = new Token( "rule", _RULE ); - rulestart->down = $0; // initial lhs is already on the stack from "rules" - $$ = new Token( "rhs", _RHS ); - //cerr << " rhs: " << $$->text << "(" << $$ << ")" << endl; - $0->left = $$; - } - | rhs lhs - { - //cerr << "rhs2: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - Token *temp = new Token( "rule", _RULE ); - rulestart->addLeftTail( temp ); - temp->down = $2; - $$ = new Token( "rhs", _RHS ); - //cerr << " rhs: " << $$->text << "(" << $$ << ")" << endl; - $2->left = $$; - } - | rhs ';' lhs - { - //cerr << "rhs3: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ") " << $3->text << "(" << $3 << ")" << endl; - $1->addDownLeftTail( $2 ); - Token *temp = new Token( "rule", _RULE ); - rulestart->addLeftTail( temp ); - temp->down = $3; - $$ = new Token( "rhs", _RHS ); - //cerr << " rhs: " << $$->text << "(" << $$ << ")" << endl; - $3->left = $$; - } - | rhs prod - { - //cerr << "rhs4: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->addDownLeftTail( $2 ); - $$ = $1; - } - | rhs '|' - { - //cerr << "rhs5: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->addDownLeftTail( $2 ); - $$ = new Token( "rhs", _RHS ); - $1->left = $$; - //cerr << " rhs: " << $$->text << "(" << $$ << ")" << endl; - } - ; - -prod : action - | IDENTIFIER - | CHARACTER - | prec - ; - -prec : PREC name - { - //cerr << "prec: " << $1->text << "(" << $1 << ") " << $2->text << "(" << $2 << ")" << endl; - $1->left = $2; - $$ = new Token( "prec", _PREC ); - $$->down = $1; - } - ; - -action : '{' - { lexC(); } - ccode_opt - { - // Remove the trailing '}' added in lex. - string temp( lexYacc() ); - $$ = new Token( temp.substr( 0, temp.length() - 1 ), CODE ); - } - '}' - { - $1->left = $4; - $4->left = $5; - $$ = new Token( "action", _ACTION ); - $$->down = $1; - } - ; - -usersection_opt : /* empty */ - { - //cerr << "usersection_opt" << endl; - // attach remaining WS to fictitious code - Token *temp = new Token( "", ws_list, CODE ); - $$ = new Token( "usersection_opt", _USERSECTION_OPT ); - $$->down = temp; - } - | MARK - { lexC(); } - ccode_opt - { - Token *temp = new Token( lexYacc(), CODE ); - //cerr << "usersection_opt: " << $1->text << " " << temp->text << endl; - $1->left = temp; - $$ = new Token( "usersection_opt", _USERSECTION_OPT ); - $$->down = $1; - } - ; - -ccode_opt : /* empty */ {} - | blocks - ; - -// This rule matches internal braces "{}" in C code to the level of the braces of a union/action. These -// internal braces are returned as Tokens from the lexer but are unused because the braces are already -// concatenated into the code string built by the lexer. Therefore, the tokens for the braces are immediately -// deleted. - -blocks : '{' { delete $1; } ccode_opt '}' { delete $4; } - | blocks '{' { delete $2; } ccode_opt '}' { delete $5; } - ; -%% - -/* Local Variables: */ -/* fill-column: 110 */ -/* compile-command: "gmake" */ -/* End: */