lex.c

本工具提供一个词法分析器和语法分析器的集成开发环境

/*
 * lex.c	--	Generate all of the lexical type files: parser.dlg tokens.h
 *
 * SOFTWARE RIGHTS
 *
 * We reserve no LEGAL rights to the Purdue Compiler Construction Tool
 * Set (PCCTS) -- PCCTS is in the public domain.  An individual or
 * company may do whatever they wish with source code distributed with
 * PCCTS or the code generated by PCCTS, including the incorporation of
 * PCCTS, or its output, into commerical software.
 *
 * We encourage users to develop software with PCCTS.  However, we do ask
 * that credit is given to us for developing PCCTS.  By "credit",
 * we mean that if you incorporate our source code into one of your
 * programs (commercial product, research project, or otherwise) that you
 * acknowledge this fact somewhere in the documentation, research report,
 * etc...  If you like PCCTS and have developed a nice tool with the
 * output, please mention that you developed it using PCCTS.  In
 * addition, we ask that this header remain intact in our source code.
 * As long as these guidelines are kept, we expect to continue enhancing
 * this system and expect to make other tools available as they are
 * completed.
 *
 * ANTLR 1.33
 * Terence Parr
 * Parr Research Corporation
 * with Purdue University and AHPCRC, University of Minnesota
 * 1989-1998
 */

#include <stdio.h>
#include <ctype.h>
/* MR1						                                                */
/* MR1  10-Apr-97  MR1	Replace use of __STDC__ with __USE_PROTOS	    */
/* MR1				                                                        */
#include "pcctscfg.h"
#include "set.h"
#include "syn.h"
#include "hash.h"
#include "generic.h"

#define DLGErrorString "invalid token"

/* Generate a complete lexical description of the lexemes found in the grammar */
void
#ifdef __USE_PROTOS
genLexDescr( void )
#else
genLexDescr( )
#endif
{
	ListNode *p;
	FILE *dlgFile = fopen(OutMetaName(DlgFileName), "w");
	require(dlgFile!=NULL, eMsg1("genLexFile: cannot open %s", OutMetaName(DlgFileName)) );
#if SPECIAL_FOPEN
	special_fopen_actions(OutMetaName(DlgFileName));	             /* MR1 */
#endif
	fprintf(dlgFile, "<<\n");
	fprintf(dlgFile, "/* %s -- DLG Description of scanner\n", DlgFileName);
	fprintf(dlgFile, " *\n");
	fprintf(dlgFile, " * Generated from:");
	{int i; for (i=0; i<NumFiles; i++) fprintf(dlgFile, " %s", FileStr[i]);}
	fprintf(dlgFile, "\n");
	fprintf(dlgFile, " *\n");
	fprintf(dlgFile, " * Terence Parr, Will Cohen, and Hank Dietz: 1989-1998\n");
	fprintf(dlgFile, " * Purdue University Electrical Engineering\n");
	fprintf(dlgFile, " * With AHPCRC, University of Minnesota\n");
	fprintf(dlgFile, " * ANTLR Version %s\n", Version);
	fprintf(dlgFile, " */\n\n");
    fprintf(dlgFile, "#define ANTLR_VERSION	%s\n", VersionDef);
    if (FirstAction != NULL ) dumpAction( FirstAction, dlgFile, 0, -1, 0, 1 );  /* MR11 */
	if ( GenCC )
	{
		if ( !UserDefdTokens ) fprintf(dlgFile, "#include \"%s\"\n", DefFileName);
		else fprintf(dlgFile, "#include %s\n", UserTokenDefsFile);
		fprintf(dlgFile, "#include \"%s\"\n", ATOKEN_H);
		if ( GenAST ) fprintf(dlgFile, "#include \"%s\"\n", ASTBASE_H);
		if ( HdrAction != NULL ) dumpAction( HdrAction, dlgFile, 0, -1, 0, 1 );
	}
	else
	{
		fprintf(dlgFile, "#include \"pcctscfg.h\"\n");
		fprintf(dlgFile, "#include PCCTS_STDIO_H\n");
		if ( strcmp(ParserName, DefaultParserName)!=0 )
			fprintf(dlgFile, "#define %s %s\n", DefaultParserName, ParserName);
		if ( strcmp(ParserName, DefaultParserName)!=0 )
			fprintf(dlgFile, "#include \"%s\"\n", RemapFileName);
		if ( HdrAction != NULL ) dumpAction( HdrAction, dlgFile, 0, -1, 0, 1 );
		if ( FoundGuessBlk )
		{
			fprintf(dlgFile, "#define ZZCAN_GUESS\n");
			fprintf(dlgFile, "#include PCCTS_SETJMP_H\n");
		}
		if ( OutputLL_k > 1 ) fprintf(dlgFile, "#define LL_K %d\n", OutputLL_k);
		if ( DemandLookahead ) fprintf(dlgFile, "#define DEMAND_LOOK\n");
		fprintf(dlgFile, "#include \"antlr.h\"\n");
		if ( GenAST ) {
			fprintf(dlgFile, "#include \"ast.h\"\n");
		}
		if ( UserDefdTokens )
			fprintf(dlgFile, "#include %s\n", UserTokenDefsFile);
		/* still need this one as it has the func prototypes */
		fprintf(dlgFile, "#include \"%s\"\n", DefFileName);
		fprintf(dlgFile, "#include \"dlgdef.h\"\n");
		fprintf(dlgFile, "LOOKAHEAD\n");
		fprintf(dlgFile, "void zzerraction()\n");
		fprintf(dlgFile, "{\n");
		fprintf(dlgFile, "\t(*zzerr)(\"%s\");\n", DLGErrorString);
		fprintf(dlgFile, "\tzzadvance();\n");
		fprintf(dlgFile, "\tzzskip();\n");
		fprintf(dlgFile, "}\n");
	}
	fprintf(dlgFile, ">>\n\n");

	/* dump all actions */

/* MR1									                                    */
/* MR1  11-Apr-97	Provide mechanism for inserting code into DLG class     */
/* MR1	 	   	  via <<%%lexmember ....>> & <<%%lexprefix ...>>            */
/* MR1				                					    */
          if (LexActions != NULL) {
            for (p = LexActions->next; p!=NULL; p=p->next)
		{
/* MR1 */	fprintf(dlgFile, "<<%%%%lexaction\n");
			dumpAction( (char *)p->elem, dlgFile, 0, -1, 0, 1 );
			fprintf(dlgFile, ">>\n\n");
		}
	  };

/* MR1 */ if (GenCC) {
/* MR1 */   fprintf(dlgFile,"<<%%%%parserclass %s>>\n\n",CurrentClassName);
/* MR1 */ };

/* MR1 */ if (LexPrefixActions != NULL) {
/* MR1 */   for (p = LexPrefixActions->next; p!=NULL; p=p->next)
/* MR1 */       {
/* MR1 */               fprintf(dlgFile, "<<%%%%lexprefix\n");
/* MR1 */               dumpAction( (char *)p->elem, dlgFile, 0, -1, 0, 1 );
/* MR1 */               fprintf(dlgFile, ">>\n\n");
/* MR1 */       }
/* MR1 */ };

/* MR1 */ if (LexMemberActions != NULL) {
/* MR1 */   for (p = LexMemberActions->next; p!=NULL; p=p->next)
/* MR1 */       {
/* MR1 */               fprintf(dlgFile, "<<%%%%lexmember\n");
/* MR1 */               dumpAction( (char *)p->elem, dlgFile, 0, -1, 0, 1 );
/* MR1 */               fprintf(dlgFile, ">>\n\n");
/* MR1 */       }
/* MR1 */ };

	/* dump all regular expression rules/actions (skip sentinel node) */
	if ( ExprOrder == NULL ) {
		warnNoFL("no regular expressions found in grammar");
	}
	else dumpLexClasses(dlgFile);
	fprintf(dlgFile, "%%%%\n");
	fclose( dlgFile );
}

/* For each lexical class, scan ExprOrder looking for expressions
 * in that lexical class.  Print out only those that match.
 * Each element of the ExprOrder list has both an expr and an lclass
 * field.
 */
void
#ifdef __USE_PROTOS
dumpLexClasses( FILE *dlgFile )
#else
dumpLexClasses( dlgFile )
FILE *dlgFile;
#endif
{
	int i;
	TermEntry *t;
	ListNode *p;
	Expr *q;

	for (i=0; i<NumLexClasses; i++)
	{
		fprintf(dlgFile, "\n%%%%%s\n\n", lclass[i].classnum);
		for (p=ExprOrder->next; p!=NULL; p=p->next)
		{
			q = (Expr *) p->elem;
			if ( q->lclass != i ) continue;
			lexmode(i);
			t = (TermEntry *) hash_get(Texpr, q->expr);
			require(t!=NULL, eMsg1("genLexDescr: rexpr %s not in hash table",q->expr) );
			if ( t->token == EpToken ) continue;
			fprintf(dlgFile, "%s\n\t<<\n", StripQuotes(q->expr));
			/* replace " killed by StripQuotes() */
			q->expr[ strlen(q->expr) ] = '"';
			if ( !GenCC ) {
				if ( TokenString(t->token) != NULL )
					fprintf(dlgFile, "\t\tNLA = %s;\n", TokenString(t->token));
				else
					fprintf(dlgFile, "\t\tNLA = %d;\n", t->token);
			}
			if ( t->action != NULL ) dumpAction( t->action, dlgFile, 2,-1,0,1 );
			if ( GenCC ) {
				if ( TokenString(t->token) != NULL )
					fprintf(dlgFile, "\t\treturn %s;\n", TokenString(t->token));
				else
					fprintf(dlgFile, "\t\treturn (ANTLRTokenType)%d;\n", t->token);
			}
			fprintf(dlgFile, "\t>>\n\n");
		}
	}
}

/* Strip the leading path (if any) from a filename */
char *
#ifdef __USE_PROTOS
StripPath( char *fileName )
#else
StripPath( fileName )
char *fileName;
#endif
{
	char *p;
	static char dirSym[2] = DirectorySymbol;

	if(NULL != (p = strrchr(fileName, dirSym[0])))
		p++;
	else
		p = fileName;

	return(p);
}

/* Generate a list of #defines && list of struct definitions for
 * aggregate retv's */
void
#ifdef __USE_PROTOS
genDefFile( void )
#else
genDefFile( )
#endif
{
	int i;

	/* If C++ mode and #tokdef used, then don't need anything in here since
	 * C++ puts all definitions in the class file name.
	 */
	if ( GenCC && UserTokenDefsFile ) return;
    if ( MR_Inhibit_Tokens_h_Gen) return;

	DefFile = fopen(OutMetaName(DefFileName), "w");
	require(DefFile!=NULL, eMsg1("genDefFile: cannot open %s", OutMetaName(DefFileName)) );
#if SPECIAL_FOPEN
	special_fopen_actions(OutMetaName(DefFileName));	             /* MR1 */
#endif
	fprintf(DefFile, "#ifndef %s\n", StripPath(gate_symbol(DefFileName)));
	fprintf(DefFile, "#define %s\n", StripPath(gate_symbol(DefFileName)));

	fprintf(DefFile, "/* %s -- List of labelled tokens and stuff\n", DefFileName);
	fprintf(DefFile, " *\n");
	fprintf(DefFile, " * Generated from:");
	for (i=0; i<NumFiles; i++) fprintf(DefFile, " %s", FileStr[i]);
	fprintf(DefFile, "\n");
	fprintf(DefFile, " *\n");
	fprintf(DefFile, " * Terence Parr, Will Cohen, and Hank Dietz: 1989-1998\n");
	fprintf(DefFile, " * Purdue University Electrical Engineering\n");
	fprintf(DefFile, " * ANTLR Version %s\n", Version);
	fprintf(DefFile, " */\n");

	if ( !GenCC && LexGen ) {
		fprintf(DefFile,"#define zzEOF_TOKEN %d\n",
				TokenInd!=NULL?TokenInd[EofToken]:EofToken);
	}

	if ( !UserDefdTokens )
	{
		int first=1;

		if ( GenCC ) fprintf(DefFile, "enum ANTLRTokenType {\n");
		for (i=1; i<TokenNum; i++)
		{
			/* Don't do EpToken or expr w/o labels */
			if ( TokenString(i)!=NULL && i != EpToken )
			{
				TermEntry *p;
				
				if ( WarningLevel>1 )
				{
					int j;
					/* look in all lexclasses for the reg expr */

/* MR10  Derek Pappas                                                */
/* MR10     A #tokclass doesn't have associated regular expressiones */
/* MR10        so don't warn user about it's omission                */

                    p = (TermEntry *) hash_get(Tname, TokenString(i));

                    if (p != NULL && ! p->classname) {
    					for (j=0; j<NumLexClasses; j++)
    					{
    						lexmode(j);
    						if ( ExprString(i)!=NULL ) break;
    					}
    					if ( j>=NumLexClasses )
    					{
    						warnNoFL(eMsg1("token label has no associated rexpr: %s",TokenString(i)));
    					}
                    };
				}
				require((p=(TermEntry *)hash_get(Tname, TokenString(i))) != NULL,
						"token not in sym tab when it should be");
				if ( !p->classname )
				{
					if ( GenCC ) {
						if ( !first ) fprintf(DefFile, ",\n");
						first = 0;
						fprintf(DefFile, "\t%s=%d", TokenString(i), i);
					}
					else
						fprintf(DefFile, "#define %s %d\n", TokenString(i), i);
				}
			}
		}
/* MR1									                                    */
/* MR1  10-Apr-97 133MR1	Prevent use of varying sizes of integer	    */
/* MR1				for the enum ANTLRTokenType                             */
/* MR1								           */
		if ( GenCC ) {				                                 /* MR1 */
 		 	 if ( !first ) fprintf(DefFile, ",\n");                  /* MR14 */
		     fprintf(DefFile, "\tDLGminToken=0");                    /* MR1 */
		     fprintf(DefFile, ",\n\tDLGmaxToken=9999};\n");          /* MR1 */
                };						                             /* MR1 */
	}

	if ( !GenCC ) GenRulePrototypes(DefFile, SynDiag);

	fprintf(DefFile, "\n#endif\n");
}

void
#ifdef __USE_PROTOS
GenRemapFile( void )
#else
GenRemapFile( )
#endif
{
	if ( strcmp(ParserName, DefaultParserName)!=0 )
	{
		FILE *f;
		int i;

		f = fopen(OutMetaName(RemapFileName), "w");
		require(f!=NULL, eMsg1("GenRemapFile: cannot open %s", OutMetaName(RemapFileName)) );
#ifdef SPECIAL_FOPEN
		special_fopen_actions(OutMetaName(RemapFileName));           /* MR1 */
#endif
		fprintf(f, "/* %s -- List of symbols to remap\n", RemapFileName);
		fprintf(f, " *\n");
		fprintf(f, " * Generated from:");
		for (i=0; i<NumFiles; i++) fprintf(f, " %s", FileStr[i]);
		fprintf(f, "\n");
		fprintf(f, " *\n");
		fprintf(f, " * Terence Parr, Will Cohen, and Hank Dietz: 1989-1998\n");
		fprintf(f, " * Purdue University Electrical Engineering\n");
		fprintf(f, " * ANTLR Version %s\n", Version);
		fprintf(f, " */\n");

		GenRuleFuncRedefs(f, SynDiag);
		GenPredefinedSymbolRedefs(f);
		if ( GenAST ) GenASTSymbolRedefs(f);
		GenSetRedefs(f);

		fclose(f);
	}
}

/* Generate a bunch of #defines that rename all functions to be "ParserName_func" */
void
#ifdef __USE_PROTOS
GenRuleFuncRedefs( FILE *f, Junction *p )
#else
GenRuleFuncRedefs( f, p )
FILE *f;
Junction *p;
#endif
{
	fprintf(f, "\n/* rename rule functions to be 'ParserName_func' */\n");
	while ( p!=NULL )
	{
		fprintf(f, "#define %s %s_%s\n", p->rname, ParserName, p->rname);
		p = (Junction *)p->p2;
	}
}

/* Generate a bunch of #defines that rename all standard symbols to be
 * "ParserName_symbol".  The list of standard symbols to change is in
 * globals.c.
 */
void
#ifdef __USE_PROTOS
GenPredefinedSymbolRedefs( FILE *f )
#else
GenPredefinedSymbolRedefs( f )
FILE *f;
#endif
{
	char **p;

	fprintf(f, "\n/* rename PCCTS-supplied symbols to be 'ParserName_symbol' */\n");
	for (p = &StandardSymbols[0]; *p!=NULL; p++)
	{
		fprintf(f, "#define %s %s_%s\n", *p, ParserName, *p);
	}
}

/* Generate a bunch of #defines that rename all AST symbols to be
 * "ParserName_symbol".  The list of AST symbols to change is in
 * globals.c.
 */
void
#ifdef __USE_PROTOS
GenASTSymbolRedefs( FILE *f )