sqlite/tool/lempar.c
drh 0bd1f4ea5a Added the %fallback directive to the lemon parser generator and used this
in the parser to make the parse tables much smaller.  This reduced the size
of the library by 15K. (CVS 605)

FossilOrigin-Name: 7ac5bd293cbb2bf252f31f1571f7efac7e77280a
2002-06-06 18:54:39 +00:00

658 lines
21 KiB
C

/* Driver template for the LEMON parser generator.
** The author disclaims copyright to this source code.
*/
/* First off, code is include which follows the "include" declaration
** in the input file. */
#include <stdio.h>
%%
/* Next is all token values, in a form suitable for use by makeheaders.
** This section will be null unless lemon is run with the -m switch.
*/
/*
** These constants (all generated automatically by the parser generator)
** specify the various kinds of tokens (terminals) that the parser
** understands.
**
** Each symbol here is a terminal symbol in the grammar.
*/
%%
/* Make sure the INTERFACE macro is defined.
*/
#ifndef INTERFACE
# define INTERFACE 1
#endif
/* The next thing included is series of defines which control
** various aspects of the generated parser.
** YYCODETYPE is the data type used for storing terminal
** and nonterminal numbers. "unsigned char" is
** used if there are fewer than 250 terminals
** and nonterminals. "int" is used otherwise.
** YYNOCODE is a number of type YYCODETYPE which corresponds
** to no legal terminal or nonterminal number. This
** number is used to fill in empty slots of the hash
** table.
** YYFALLBACK If defined, this indicates that one or more tokens
** have fall-back values which should be used if the
** original value of the token will not parse.
** YYACTIONTYPE is the data type used for storing terminal
** and nonterminal numbers. "unsigned char" is
** used if there are fewer than 250 rules and
** states combined. "int" is used otherwise.
** ParseTOKENTYPE is the data type used for minor tokens given
** directly to the parser from the tokenizer.
** YYMINORTYPE is the data type used for all minor tokens.
** This is typically a union of many types, one of
** which is ParseTOKENTYPE. The entry in the union
** for base tokens is called "yy0".
** YYSTACKDEPTH is the maximum depth of the parser's stack.
** ParseARG_SDECL A static variable declaration for the %extra_argument
** ParseARG_PDECL A parameter declaration for the %extra_argument
** ParseARG_STORE Code to store %extra_argument into yypParser
** ParseARG_FETCH Code to extract %extra_argument from yypParser
** YYNSTATE the combined number of states.
** YYNRULE the number of rules in the grammar
** YYERRORSYMBOL is the code number of the error symbol. If not
** defined, then do no error processing.
*/
%%
#define YY_NO_ACTION (YYNSTATE+YYNRULE+2)
#define YY_ACCEPT_ACTION (YYNSTATE+YYNRULE+1)
#define YY_ERROR_ACTION (YYNSTATE+YYNRULE)
/* Next is the action table. Each entry in this table contains
**
** + An integer which is the number representing the look-ahead
** token
**
** + An integer indicating what action to take. Number (N) between
** 0 and YYNSTATE-1 mean shift the look-ahead and go to state N.
** Numbers between YYNSTATE and YYNSTATE+YYNRULE-1 mean reduce by
** rule N-YYNSTATE. Number YYNSTATE+YYNRULE means that a syntax
** error has occurred. Number YYNSTATE+YYNRULE+1 means the parser
** accepts its input.
**
** + A pointer to the next entry with the same hash value.
**
** The action table is really a series of hash tables. Each hash
** table contains a number of entries which is a power of two. The
** "state" table (which follows) contains information about the starting
** point and size of each hash table.
*/
struct yyActionEntry {
YYCODETYPE lookahead; /* The value of the look-ahead token */
YYCODETYPE next; /* Next entry + 1. Zero at end of collision chain */
YYACTIONTYPE action; /* Action to take for this look-ahead */
};
typedef struct yyActionEntry yyActionEntry;
static const yyActionEntry yyActionTable[] = {
%%
};
/* The state table contains information needed to look up the correct
** action in the action table, given the current state of the parser.
** Information needed includes:
**
** + A pointer to the start of the action hash table in yyActionTable.
**
** + The number of entries in the action hash table.
**
** + The default action. This is the action to take if no entry for
** the given look-ahead is found in the action hash table.
*/
struct yyStateEntry {
const yyActionEntry *hashtbl; /* Start of the hash table in yyActionTable */
YYCODETYPE nEntry; /* Number of entries in action hash table */
YYACTIONTYPE actionDefault; /* Default action if look-ahead not found */
};
typedef struct yyStateEntry yyStateEntry;
static const yyStateEntry yyStateTable[] = {
%%
};
/* The next table maps tokens into fallback tokens. If a construct
** like the following:
**
** %fallback ID X Y Z.
**
** appears in the grammer, then ID becomes a fallback token for X, Y,
** and Z. Whenever one of the tokens X, Y, or Z is input to the parser
** but it does not parse, the type of the token is changed to ID and
** the parse is retried before an error is thrown.
*/
#ifdef YYFALLBACK
static const YYCODETYPE yyFallback[] = {
%%
};
#endif /* YYFALLBACK */
/* The following structure represents a single element of the
** parser's stack. Information stored includes:
**
** + The state number for the parser at this level of the stack.
**
** + The value of the token stored at this level of the stack.
** (In other words, the "major" token.)
**
** + The semantic value stored at this level of the stack. This is
** the information used by the action routines in the grammar.
** It is sometimes called the "minor" token.
*/
struct yyStackEntry {
int stateno; /* The state-number */
int major; /* The major token value. This is the code
** number for the token at this stack level */
YYMINORTYPE minor; /* The user-supplied minor token value. This
** is the value of the token */
};
typedef struct yyStackEntry yyStackEntry;
/* The state of the parser is completely contained in an instance of
** the following structure */
struct yyParser {
int yyidx; /* Index of top element in stack */
int yyerrcnt; /* Shifts left before out of the error */
yyStackEntry *yytop; /* Pointer to the top stack element */
ParseARG_SDECL /* A place to hold %extra_argument */
yyStackEntry yystack[YYSTACKDEPTH]; /* The parser's stack */
};
typedef struct yyParser yyParser;
#ifndef NDEBUG
#include <stdio.h>
static FILE *yyTraceFILE = 0;
static char *yyTracePrompt = 0;
#endif /* NDEBUG */
#ifndef NDEBUG
/*
** Turn parser tracing on by giving a stream to which to write the trace
** and a prompt to preface each trace message. Tracing is turned off
** by making either argument NULL
**
** Inputs:
** <ul>
** <li> A FILE* to which trace output should be written.
** If NULL, then tracing is turned off.
** <li> A prefix string written at the beginning of every
** line of trace output. If NULL, then tracing is
** turned off.
** </ul>
**
** Outputs:
** None.
*/
void ParseTrace(FILE *TraceFILE, char *zTracePrompt){
yyTraceFILE = TraceFILE;
yyTracePrompt = zTracePrompt;
if( yyTraceFILE==0 ) yyTracePrompt = 0;
else if( yyTracePrompt==0 ) yyTraceFILE = 0;
}
#endif /* NDEBUG */
#ifndef NDEBUG
/* For tracing shifts, the names of all terminals and nonterminals
** are required. The following table supplies these names */
static const char *yyTokenName[] = {
%%
};
#endif /* NDEBUG */
#ifndef NDEBUG
/* For tracing reduce actions, the names of all rules are required.
*/
static const char *yyRuleName[] = {
%%
};
#endif /* NDEBUG */
/*
** This function returns the symbolic name associated with a token
** value.
*/
const char *ParseTokenName(int tokenType){
#ifndef NDEBUG
if( tokenType>0 && tokenType<(sizeof(yyTokenName)/sizeof(yyTokenName[0])) ){
return yyTokenName[tokenType];
}else{
return "Unknown";
}
#else
return "";
#endif
}
/*
** This function allocates a new parser.
** The only argument is a pointer to a function which works like
** malloc.
**
** Inputs:
** A pointer to the function used to allocate memory.
**
** Outputs:
** A pointer to a parser. This pointer is used in subsequent calls
** to Parse and ParseFree.
*/
void *ParseAlloc(void *(*mallocProc)(size_t)){
yyParser *pParser;
pParser = (yyParser*)(*mallocProc)( (size_t)sizeof(yyParser) );
if( pParser ){
pParser->yyidx = -1;
}
return pParser;
}
/* The following function deletes the value associated with a
** symbol. The symbol can be either a terminal or nonterminal.
** "yymajor" is the symbol code, and "yypminor" is a pointer to
** the value.
*/
static void yy_destructor(YYCODETYPE yymajor, YYMINORTYPE *yypminor){
switch( yymajor ){
/* Here is inserted the actions which take place when a
** terminal or non-terminal is destroyed. This can happen
** when the symbol is popped from the stack during a
** reduce or during error processing or when a parser is
** being destroyed before it is finished parsing.
**
** Note: during a reduce, the only symbols destroyed are those
** which appear on the RHS of the rule, but which are not used
** inside the C code.
*/
%%
default: break; /* If no destructor action specified: do nothing */
}
}
/*
** Pop the parser's stack once.
**
** If there is a destructor routine associated with the token which
** is popped from the stack, then call it.
**
** Return the major token number for the symbol popped.
*/
static int yy_pop_parser_stack(yyParser *pParser){
YYCODETYPE yymajor;
if( pParser->yyidx<0 ) return 0;
#ifndef NDEBUG
if( yyTraceFILE && pParser->yyidx>=0 ){
fprintf(yyTraceFILE,"%sPopping %s\n",
yyTracePrompt,
yyTokenName[pParser->yytop->major]);
}
#endif
yymajor = pParser->yytop->major;
yy_destructor( yymajor, &pParser->yytop->minor);
pParser->yyidx--;
pParser->yytop--;
return yymajor;
}
/*
** Deallocate and destroy a parser. Destructors are all called for
** all stack elements before shutting the parser down.
**
** Inputs:
** <ul>
** <li> A pointer to the parser. This should be a pointer
** obtained from ParseAlloc.
** <li> A pointer to a function used to reclaim memory obtained
** from malloc.
** </ul>
*/
void ParseFree(
void *p, /* The parser to be deleted */
void (*freeProc)(void*) /* Function used to reclaim memory */
){
yyParser *pParser = (yyParser*)p;
if( pParser==0 ) return;
while( pParser->yyidx>=0 ) yy_pop_parser_stack(pParser);
(*freeProc)((void*)pParser);
}
/*
** Find the appropriate action for a parser given the look-ahead token.
**
** If the look-ahead token is YYNOCODE, then check to see if the action is
** independent of the look-ahead. If it is, return the action, otherwise
** return YY_NO_ACTION.
*/
static int yy_find_parser_action(
yyParser *pParser, /* The parser */
int iLookAhead /* The look-ahead token */
){
const yyStateEntry *pState; /* Appropriate entry in the state table */
const yyActionEntry *pAction; /* Action appropriate for the look-ahead */
int iFallback; /* Fallback token */
/* if( pParser->yyidx<0 ) return YY_NO_ACTION; */
pState = &yyStateTable[pParser->yytop->stateno];
if( pState->nEntry==0 ){
return pState->actionDefault;
}else if( iLookAhead!=YYNOCODE ){
pAction = &pState->hashtbl[iLookAhead % pState->nEntry];
while( 1 ){
if( pAction->lookahead==iLookAhead ) return pAction->action;
if( pAction->next==0 ) break;
pAction = &pState->hashtbl[pAction->next-1];
}
#ifdef YYFALLBACK
if( iLookAhead<sizeof(yyFallback)/sizeof(yyFallback[0])
&& (iFallback = yyFallback[iLookAhead])!=0 ){
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE, "%sFALLBACK %s => %s\n",
yyTracePrompt, yyTokenName[iLookAhead], yyTokenName[iFallback]);
}
#endif
return yy_find_parser_action(pParser, iFallback);
}
#endif
}else if( pState->hashtbl->lookahead!=YYNOCODE ){
return YY_NO_ACTION;
}
return pState->actionDefault;
}
/*
** Perform a shift action.
*/
static void yy_shift(
yyParser *yypParser, /* The parser to be shifted */
int yyNewState, /* The new state to shift in */
int yyMajor, /* The major token to shift in */
YYMINORTYPE *yypMinor /* Pointer ot the minor token to shift in */
){
yypParser->yyidx++;
yypParser->yytop++;
if( yypParser->yyidx>=YYSTACKDEPTH ){
ParseARG_FETCH;
yypParser->yyidx--;
yypParser->yytop--;
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sStack Overflow!\n",yyTracePrompt);
}
#endif
while( yypParser->yyidx>=0 ) yy_pop_parser_stack(yypParser);
/* Here code is inserted which will execute if the parser
** stack every overflows */
%%
ParseARG_STORE; /* Suppress warning about unused %extra_argument var */
return;
}
yypParser->yytop->stateno = yyNewState;
yypParser->yytop->major = yyMajor;
yypParser->yytop->minor = *yypMinor;
#ifndef NDEBUG
if( yyTraceFILE && yypParser->yyidx>0 ){
int i;
fprintf(yyTraceFILE,"%sShift %d\n",yyTracePrompt,yyNewState);
fprintf(yyTraceFILE,"%sStack:",yyTracePrompt);
for(i=1; i<=yypParser->yyidx; i++)
fprintf(yyTraceFILE," %s",yyTokenName[yypParser->yystack[i].major]);
fprintf(yyTraceFILE,"\n");
}
#endif
}
/* The following table contains information about every rule that
** is used during the reduce.
*/
static struct {
YYCODETYPE lhs; /* Symbol on the left-hand side of the rule */
unsigned char nrhs; /* Number of right-hand side symbols in the rule */
} yyRuleInfo[] = {
%%
};
static void yy_accept(yyParser*); /* Forward Declaration */
/*
** Perform a reduce action and the shift that must immediately
** follow the reduce.
*/
static void yy_reduce(
yyParser *yypParser, /* The parser */
int yyruleno /* Number of the rule by which to reduce */
){
int yygoto; /* The next state */
int yyact; /* The next action */
YYMINORTYPE yygotominor; /* The LHS of the rule reduced */
yyStackEntry *yymsp; /* The top of the parser's stack */
int yysize; /* Amount to pop the stack */
ParseARG_FETCH;
yymsp = yypParser->yytop;
#ifndef NDEBUG
if( yyTraceFILE && yyruleno>=0
&& yyruleno<sizeof(yyRuleName)/sizeof(yyRuleName[0]) ){
fprintf(yyTraceFILE, "%sReduce [%s].\n", yyTracePrompt,
yyRuleName[yyruleno]);
}
#endif /* NDEBUG */
switch( yyruleno ){
/* Beginning here are the reduction cases. A typical example
** follows:
** case 0:
** #line <lineno> <grammarfile>
** { ... } // User supplied code
** #line <lineno> <thisfile>
** break;
*/
%%
};
yygoto = yyRuleInfo[yyruleno].lhs;
yysize = yyRuleInfo[yyruleno].nrhs;
yypParser->yyidx -= yysize;
yypParser->yytop -= yysize;
yyact = yy_find_parser_action(yypParser,yygoto);
if( yyact < YYNSTATE ){
yy_shift(yypParser,yyact,yygoto,&yygotominor);
}else if( yyact == YYNSTATE + YYNRULE + 1 ){
yy_accept(yypParser);
}
}
/*
** The following code executes when the parse fails
*/
static void yy_parse_failed(
yyParser *yypParser /* The parser */
){
ParseARG_FETCH;
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sFail!\n",yyTracePrompt);
}
#endif
while( yypParser->yyidx>=0 ) yy_pop_parser_stack(yypParser);
/* Here code is inserted which will be executed whenever the
** parser fails */
%%
ParseARG_STORE; /* Suppress warning about unused %extra_argument variable */
}
/*
** The following code executes when a syntax error first occurs.
*/
static void yy_syntax_error(
yyParser *yypParser, /* The parser */
int yymajor, /* The major type of the error token */
YYMINORTYPE yyminor /* The minor type of the error token */
){
ParseARG_FETCH;
#define TOKEN (yyminor.yy0)
%%
ParseARG_STORE; /* Suppress warning about unused %extra_argument variable */
}
/*
** The following is executed when the parser accepts
*/
static void yy_accept(
yyParser *yypParser /* The parser */
){
ParseARG_FETCH;
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sAccept!\n",yyTracePrompt);
}
#endif
while( yypParser->yyidx>=0 ) yy_pop_parser_stack(yypParser);
/* Here code is inserted which will be executed whenever the
** parser accepts */
%%
ParseARG_STORE; /* Suppress warning about unused %extra_argument variable */
}
/* The main parser program.
** The first argument is a pointer to a structure obtained from
** "ParseAlloc" which describes the current state of the parser.
** The second argument is the major token number. The third is
** the minor token. The fourth optional argument is whatever the
** user wants (and specified in the grammar) and is available for
** use by the action routines.
**
** Inputs:
** <ul>
** <li> A pointer to the parser (an opaque structure.)
** <li> The major token number.
** <li> The minor token number.
** <li> An option argument of a grammar-specified type.
** </ul>
**
** Outputs:
** None.
*/
void Parse(
void *yyp, /* The parser */
int yymajor, /* The major token code number */
ParseTOKENTYPE yyminor /* The value for the token */
ParseARG_PDECL /* Optional %extra_argument parameter */
){
YYMINORTYPE yyminorunion;
int yyact; /* The parser action. */
int yyendofinput; /* True if we are at the end of input */
int yyerrorhit = 0; /* True if yymajor has invoked an error */
yyParser *yypParser; /* The parser */
/* (re)initialize the parser, if necessary */
yypParser = (yyParser*)yyp;
if( yypParser->yyidx<0 ){
if( yymajor==0 ) return;
yypParser->yyidx = 0;
yypParser->yyerrcnt = -1;
yypParser->yytop = &yypParser->yystack[0];
yypParser->yytop->stateno = 0;
yypParser->yytop->major = 0;
}
yyminorunion.yy0 = yyminor;
yyendofinput = (yymajor==0);
ParseARG_STORE;
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sInput %s\n",yyTracePrompt,yyTokenName[yymajor]);
}
#endif
do{
yyact = yy_find_parser_action(yypParser,yymajor);
if( yyact<YYNSTATE ){
yy_shift(yypParser,yyact,yymajor,&yyminorunion);
yypParser->yyerrcnt--;
if( yyendofinput && yypParser->yyidx>=0 ){
yymajor = 0;
}else{
yymajor = YYNOCODE;
}
}else if( yyact < YYNSTATE + YYNRULE ){
yy_reduce(yypParser,yyact-YYNSTATE);
}else if( yyact == YY_ERROR_ACTION ){
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sSyntax Error!\n",yyTracePrompt);
}
#endif
#ifdef YYERRORSYMBOL
/* A syntax error has occurred.
** The response to an error depends upon whether or not the
** grammar defines an error token "ERROR".
**
** This is what we do if the grammar does define ERROR:
**
** * Call the %syntax_error function.
**
** * Begin popping the stack until we enter a state where
** it is legal to shift the error symbol, then shift
** the error symbol.
**
** * Set the error count to three.
**
** * Begin accepting and shifting new tokens. No new error
** processing will occur until three tokens have been
** shifted successfully.
**
*/
if( yypParser->yyerrcnt<0 ){
yy_syntax_error(yypParser,yymajor,yyminorunion);
}
if( yypParser->yytop->major==YYERRORSYMBOL || yyerrorhit ){
#ifndef NDEBUG
if( yyTraceFILE ){
fprintf(yyTraceFILE,"%sDiscard input token %s\n",
yyTracePrompt,yyTokenName[yymajor]);
}
#endif
yy_destructor(yymajor,&yyminorunion);
yymajor = YYNOCODE;
}else{
while(
yypParser->yyidx >= 0 &&
yypParser->yytop->major != YYERRORSYMBOL &&
(yyact = yy_find_parser_action(yypParser,YYERRORSYMBOL)) >= YYNSTATE
){
yy_pop_parser_stack(yypParser);
}
if( yypParser->yyidx < 0 || yymajor==0 ){
yy_destructor(yymajor,&yyminorunion);
yy_parse_failed(yypParser);
yymajor = YYNOCODE;
}else if( yypParser->yytop->major!=YYERRORSYMBOL ){
YYMINORTYPE u2;
u2.YYERRSYMDT = 0;
yy_shift(yypParser,yyact,YYERRORSYMBOL,&u2);
}
}
yypParser->yyerrcnt = 3;
yyerrorhit = 1;
#else /* YYERRORSYMBOL is not defined */
/* This is what we do if the grammar does not define ERROR:
**
** * Report an error message, and throw away the input token.
**
** * If the input token is $, then fail the parse.
**
** As before, subsequent error messages are suppressed until
** three input tokens have been successfully shifted.
*/
if( yypParser->yyerrcnt<=0 ){
yy_syntax_error(yypParser,yymajor,yyminorunion);
}
yypParser->yyerrcnt = 3;
yy_destructor(yymajor,&yyminorunion);
if( yyendofinput ){
yy_parse_failed(yypParser);
}
yymajor = YYNOCODE;
#endif
}else{
yy_accept(yypParser);
yymajor = YYNOCODE;
}
}while( yymajor!=YYNOCODE && yypParser->yyidx>=0 );
return;
}