parse.y

flex 词法分析工具 类似于lex 此版本为较早前的版本

			trlcontxt = true;

			eps = mkstate( SYM_EPSILON );
			$$ = link_machines( $1,
				link_machines( eps, mkstate( '\n' ) ) );
			}

		|  re
			{
			$$ = $1;

			if ( trlcontxt )
				{
				if ( lex_compat || (varlength && headcnt == 0) )
					/* Both head and trail are
					 * variable-length.
					 */
					variable_trail_rule = true;
				else
					trailcnt = rulelen;
				}
			}
		;


re		:  re '|' series
			{
			varlength = true;
			$$ = mkor( $1, $3 );
			}

		|  series
			{ $$ = $1; }
		;


re2		:  re '/'
			{
			/* This rule is written separately so the
			 * reduction will occur before the trailing
			 * series is parsed.
			 */

			if ( trlcontxt )
				synerr( "trailing context used twice" );
			else
				trlcontxt = true;

			if ( varlength )
				/* We hope the trailing context is
				 * fixed-length.
				 */
				varlength = false;
			else
				headcnt = rulelen;

			rulelen = 0;

			current_state_type = STATE_TRAILING_CONTEXT;
			$$ = $1;
			}
		;

series		:  series singleton
			{
			/* This is where concatenation of adjacent patterns
			 * gets done.
			 */
			$$ = link_machines( $1, $2 );
			}

		|  singleton
			{ $$ = $1; }
		;

singleton	:  singleton '*'
			{
			varlength = true;

			$$ = mkclos( $1 );
			}

		|  singleton '+'
			{
			varlength = true;
			$$ = mkposcl( $1 );
			}

		|  singleton '?'
			{
			varlength = true;
			$$ = mkopt( $1 );
			}

		|  singleton '{' NUMBER ',' NUMBER '}'
			{
			varlength = true;

			if ( $3 > $5 || $3 < 0 )
				{
				synerr( "bad iteration values" );
				$$ = $1;
				}
			else
				{
				if ( $3 == 0 )
					{
					if ( $5 <= 0 )
						{
						synerr(
						"bad iteration values" );
						$$ = $1;
						}
					else
						$$ = mkopt(
							mkrep( $1, 1, $5 ) );
					}
				else
					$$ = mkrep( $1, $3, $5 );
				}
			}

		|  singleton '{' NUMBER ',' '}'
			{
			varlength = true;

			if ( $3 <= 0 )
				{
				synerr( "iteration value must be positive" );
				$$ = $1;
				}

			else
				$$ = mkrep( $1, $3, INFINITY );
			}

		|  singleton '{' NUMBER '}'
			{
			/* The singleton could be something like "(foo)",
			 * in which case we have no idea what its length
			 * is, so we punt here.
			 */
			varlength = true;

			if ( $3 <= 0 )
				{
				synerr( "iteration value must be positive" );
				$$ = $1;
				}

			else
				$$ = link_machines( $1,
						copysingl( $1, $3 - 1 ) );
			}

		|  '.'
			{
			if ( ! madeany )
				{
				/* Create the '.' character class. */
				anyccl = cclinit();
				ccladd( anyccl, '\n' );
				cclnegate( anyccl );

				if ( useecs )
					mkeccl( ccltbl + cclmap[anyccl],
						ccllen[anyccl], nextecm,
						ecgroup, csize, csize );

				madeany = true;
				}

			++rulelen;

			$$ = mkstate( -anyccl );
			}

		|  fullccl
			{
			if ( ! cclsorted )
				/* Sort characters for fast searching.  We
				 * use a shell sort since this list could
				 * be large.
				 */
				cshell( ccltbl + cclmap[$1], ccllen[$1], true );

			if ( useecs )
				mkeccl( ccltbl + cclmap[$1], ccllen[$1],
					nextecm, ecgroup, csize, csize );

			++rulelen;

			$$ = mkstate( -$1 );
			}

		|  PREVCCL
			{
			++rulelen;

			$$ = mkstate( -$1 );
			}

		|  '"' string '"'
			{ $$ = $2; }

		|  '(' re ')'
			{ $$ = $2; }

		|  CHAR
			{
			++rulelen;

			if ( caseins && $1 >= 'A' && $1 <= 'Z' )
				$1 = clower( $1 );

			$$ = mkstate( $1 );
			}
		;

fullccl		:  '[' ccl ']'
			{ $$ = $2; }

		|  '[' '^' ccl ']'
			{
			cclnegate( $3 );
			$$ = $3;
			}
		;

ccl		:  ccl CHAR '-' CHAR
			{
			if ( caseins )
				{
				if ( $2 >= 'A' && $2 <= 'Z' )
					$2 = clower( $2 );
				if ( $4 >= 'A' && $4 <= 'Z' )
					$4 = clower( $4 );
				}

			if ( $2 > $4 )
				synerr( "negative range in character class" );

			else
				{
				for ( i = $2; i <= $4; ++i )
					ccladd( $1, i );

				/* Keep track if this ccl is staying in
				 * alphabetical order.
				 */
				cclsorted = cclsorted && ($2 > lastchar);
				lastchar = $4;
				}

			$$ = $1;
			}

		|  ccl CHAR
			{
			if ( caseins && $2 >= 'A' && $2 <= 'Z' )
				$2 = clower( $2 );

			ccladd( $1, $2 );
			cclsorted = cclsorted && ($2 > lastchar);
			lastchar = $2;
			$$ = $1;
			}

		|  ccl ccl_expr
			{
			/* Too hard to properly maintain cclsorted. */
			cclsorted = false;
			$$ = $1;
			}

		|
			{
			cclsorted = true;
			lastchar = 0;
			currccl = $$ = cclinit();
			}
		;

ccl_expr:	   CCE_ALNUM	{ CCL_EXPR(isalnum) }
		|  CCE_ALPHA	{ CCL_EXPR(isalpha) }
		|  CCE_BLANK	{ CCL_EXPR(IS_BLANK) }
		|  CCE_CNTRL	{ CCL_EXPR(iscntrl) }
		|  CCE_DIGIT	{ CCL_EXPR(isdigit) }
		|  CCE_GRAPH	{ CCL_EXPR(isgraph) }
		|  CCE_LOWER	{ CCL_EXPR(islower) }
		|  CCE_PRINT	{ CCL_EXPR(isprint) }
		|  CCE_PUNCT	{ CCL_EXPR(ispunct) }
		|  CCE_SPACE	{ CCL_EXPR(isspace) }
		|  CCE_UPPER	{
				if ( caseins )
					CCL_EXPR(islower)
				else
					CCL_EXPR(isupper)
				}
		|  CCE_XDIGIT	{ CCL_EXPR(isxdigit) }
		;
		
string		:  string CHAR
			{
			if ( caseins && $2 >= 'A' && $2 <= 'Z' )
				$2 = clower( $2 );

			++rulelen;

			$$ = link_machines( $1, mkstate( $2 ) );
			}

		|
			{ $$ = mkstate( SYM_EPSILON ); }
		;

%%


/* build_eof_action - build the "<<EOF>>" action for the active start
 *                    conditions
 */

void build_eof_action()
	{
	register int i;
	char action_text[MAXLINE];

	for ( i = 1; i <= scon_stk_ptr; ++i )
		{
		if ( sceof[scon_stk[i]] )
			format_pinpoint_message(
				"multiple <<EOF>> rules for start condition %s",
				scname[scon_stk[i]] );

		else
			{
			sceof[scon_stk[i]] = true;
			sprintf( action_text, "case YY_STATE_EOF(%s):\n",
				scname[scon_stk[i]] );
			add_action( action_text );
			}
		}

	line_directive_out( (FILE *) 0, 1 );

	/* This isn't a normal rule after all - don't count it as
	 * such, so we don't have any holes in the rule numbering
	 * (which make generating "rule can never match" warnings
	 * more difficult.
	 */
	--num_rules;
	++num_eof_rules;
	}


/* format_synerr - write out formatted syntax error */

void format_synerr( msg, arg )
char msg[], arg[];
	{
	char errmsg[MAXLINE];

	(void) sprintf( errmsg, msg, arg );
	synerr( errmsg );
	}


/* synerr - report a syntax error */

void synerr( str )
char str[];
	{
	syntaxerror = true;
	pinpoint_message( str );
	}


/* format_warn - write out formatted warning */

void format_warn( msg, arg )
char msg[], arg[];
	{
	char warn_msg[MAXLINE];

	(void) sprintf( warn_msg, msg, arg );
	warn( warn_msg );
	}


/* warn - report a warning, unless -w was given */

void warn( str )
char str[];
	{
	line_warning( str, linenum );
	}

/* format_pinpoint_message - write out a message formatted with one string,
 *			     pinpointing its location
 */

void format_pinpoint_message( msg, arg )
char msg[], arg[];
	{
	char errmsg[MAXLINE];

	(void) sprintf( errmsg, msg, arg );
	pinpoint_message( errmsg );
	}


/* pinpoint_message - write out a message, pinpointing its location */

void pinpoint_message( str )
char str[];
	{
	line_pinpoint( str, linenum );
	}


/* line_warning - report a warning at a given line, unless -w was given */

void line_warning( str, line )
char str[];
int line;
	{
	char warning[MAXLINE];

	if ( ! nowarn )
		{
		sprintf( warning, "warning, %s", str );
		line_pinpoint( warning, line );
		}
	}


/* line_pinpoint - write out a message, pinpointing it at the given line */

void line_pinpoint( str, line )
char str[];
int line;
	{
	fprintf( stderr, "\"%s\", line %d: %s\n", infilename, line, str );
	}


/* yyerror - eat up an error message from the parser;
 *	     currently, messages are ignore
 */

void yyerror( msg )
char msg[];
	{
	}