php-src/ext/standard/url_scanner.c

/* Generated by re2c 0.5 on Sun Sep 12 00:56:23 1999 */
#line 1 "xx.re"

#include <sys/types.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#define YYCTYPE char
#define YYCURSOR state->crs
#define YYLIMIT state->end
#define YYMARKER state->ptr
#define YYFILL(n)

typedef enum {
	INITIAL,
	REF
} state;

typedef struct {
	state state;
	const char *crs;
	const char *end;
	const char *ptr;
	const char *start;
	char *target;
	size_t targetsize;
	const char *data;
} lexdata;

#define FINISH { catchup(state); goto finish; }

#define BEGIN(x) \
		switch(state->state) { \
			case INITIAL: \
				catchup(state); \
				break; \
			case REF: \
				screw_url(state); \
				break; \
		} \
		state->state = x; \
		state->start = state->crs; \
		goto nextiter

#define ATTACH(s, n) { \
	size_t _newlen = state->targetsize + n; \
	state->target = realloc(state->target, _newlen + 1); \
	memcpy(state->target + state->targetsize, s, n); \
	state->targetsize = _newlen; \
	state->target[_newlen] = '\0'; \
	}
	
#define URLLEN 512
	
static void screw_url(lexdata *state)
{
	char *url;
	int len;
	char buf[URLLEN];
	const char *p, *q;
	char c;

	/* search outer limits for URI */
	for(p = state->start; p < state->crs && (c = *p); p++)
		if(c != '"' && c != ' ') break;

	/*  
	 *  we look at q-1, because q points to the character behind the last
	 *  character we are going to copy and the decision is based on that last
	 *  character 
	 */

	for(q = state->crs; q > state->start && (c = *(q-1)); q--)
		if(c != '"' && c != ' ') break;

	/* attach beginning */
	
	ATTACH(state->start, p-state->start);
	
	/* copy old URI */
	len = q - p;
	url = malloc(len + 1);
	memcpy(url, p, len);
	url[len] = '\0';
	
	/* construct new URI */
	len = snprintf(buf, sizeof(buf), "%s%c%s", url,
			memchr(state->start, '?', len) ? '&' : '?',
			state->data);
	free(url);

	/* attach new URI */
	ATTACH(buf, len);
	
	/* attach rest */
	ATTACH(q, state->crs - q);
}

static void catchup(lexdata *state) 
{
	ATTACH(state->start, (state->crs - state->start));
}

#line 115


static void url_scanner(lexdata *state)
{
	while(state->crs < state->end) {
	
	switch(state->state) {
		case INITIAL: 
{
	YYCTYPE yych;
	unsigned int yyaccept;
	goto yy0;
yy1:	++YYCURSOR;
yy0:
	if((YYLIMIT - YYCURSOR) < 7) YYFILL(7);
	yych = *YYCURSOR;
	switch(yych){
	case '\000':	goto yy7;
	case '<':	goto yy2;
	default:	goto yy4;
	}
yy2:	yych = *++YYCURSOR;
	switch(yych){
	case 'A':	case 'a':	goto yy9;
	case 'F':	case 'f':	goto yy10;
	default:	goto yy3;
	}
yy3:yy4:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy5:	switch(yych){
	case '\000':	case '<':	goto yy6;
	default:	goto yy4;
	}
yy6:
#line 126
	{ BEGIN(INITIAL); }
yy7:	yych = *++YYCURSOR;
yy8:
#line 127
	{ FINISH; }
yy9:	yych = *++YYCURSOR;
	switch(yych){
	case 'H':	case 'h':	goto yy3;
	default:	goto yy25;
	}
yy10:	yych = *++YYCURSOR;
	switch(yych){
	case 'R':	case 'r':	goto yy11;
	default:	goto yy3;
	}
yy11:	yych = *++YYCURSOR;
	switch(yych){
	case 'A':	case 'a':	goto yy12;
	default:	goto yy3;
	}
yy12:	yych = *++YYCURSOR;
	switch(yych){
	case 'M':	case 'm':	goto yy13;
	default:	goto yy3;
	}
yy13:	yych = *++YYCURSOR;
	switch(yych){
	case 'E':	case 'e':	goto yy14;
	default:	goto yy3;
	}
yy14:	yych = *++YYCURSOR;
	switch(yych){
	case 'S':	case 's':	goto yy3;
	default:	goto yy16;
	}
yy15:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy16:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy15;
	case 'S':	case 's':	goto yy17;
	default:	goto yy3;
	}
yy17:	yych = *++YYCURSOR;
	switch(yych){
	case 'R':	case 'r':	goto yy18;
	default:	goto yy3;
	}
yy18:	yych = *++YYCURSOR;
	switch(yych){
	case 'C':	case 'c':	goto yy19;
	default:	goto yy3;
	}
yy19:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy20:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy19;
	case '=':	goto yy21;
	default:	goto yy3;
	}
yy21:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy22:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy21;
	default:	goto yy23;
	}
yy23:
#line 124
	{ BEGIN(REF); }
yy24:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy25:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy24;
	case 'H':	case 'h':	goto yy26;
	default:	goto yy3;
	}
yy26:	yych = *++YYCURSOR;
	switch(yych){
	case 'R':	case 'r':	goto yy27;
	default:	goto yy3;
	}
yy27:	yych = *++YYCURSOR;
	switch(yych){
	case 'E':	case 'e':	goto yy28;
	default:	goto yy3;
	}
yy28:	yych = *++YYCURSOR;
	switch(yych){
	case 'F':	case 'f':	goto yy29;
	default:	goto yy3;
	}
yy29:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy30:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy29;
	case '=':	goto yy31;
	default:	goto yy3;
	}
yy31:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy32:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy31;
	default:	goto yy33;
	}
yy33:
#line 125
	{ BEGIN(REF); }
}
#line 128

			break;
		case REF: 
{
	YYCTYPE yych;
	unsigned int yyaccept;
	goto yy34;
yy35:	++YYCURSOR;
yy34:
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
	switch(yych){
	case '\000':	case '#':	case '>':	goto yy36;
	case '\t':	case '\v':
	case '\f':	case ' ':	case '"':	goto yy37;
	case ':':	goto yy42;
	default:	goto yy39;
	}
yy36:yy37:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy38:	switch(yych){
	case '\000':	case '"':
	case '#':	case '>':	goto yy36;
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy37;
	case ':':	goto yy42;
	default:	goto yy39;
	}
yy39:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy40:	switch(yych){
	case '\000':	case '>':	goto yy41;
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy48;
	case '"':	goto yy50;
	case '#':	goto yy51;
	case ':':	goto yy42;
	default:	goto yy39;
	}
yy41:
#line 132
	{ BEGIN(INITIAL); }
yy42:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy43:	switch(yych){
	case '\000':	case '#':	case '>':	goto yy44;
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy45;
	case '"':	goto yy47;
	default:	goto yy42;
	}
yy44:
#line 134
	{ 
			/* don't modify absolute links */
			state->state = INITIAL; BEGIN(INITIAL); 
	}
yy45:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy46:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy45;
	case '"':	goto yy47;
	default:	goto yy44;
	}
yy47:	yych = *++YYCURSOR;
	goto yy44;
yy48:	++YYCURSOR;
	if(YYLIMIT == YYCURSOR) YYFILL(1);
	yych = *YYCURSOR;
yy49:	switch(yych){
	case '\t':	case '\v':
	case '\f':	case ' ':	goto yy48;
	case '"':	goto yy50;
	default:	goto yy41;
	}
yy50:	yych = *++YYCURSOR;
	goto yy41;
yy51:	yych = *++YYCURSOR;
yy52:	YYCURSOR -= 1;
#line 133
	{ BEGIN(INITIAL); }
}
#line 138

  			break;
	}
nextiter:
	}
finish:
}

char *url_adapt(const char *src, size_t srclen, const char *data, size_t *newlen)
{
	lexdata state;

	state.state = INITIAL;
	state.start = state.crs = src;
	state.end = src + srclen;
	state.ptr = NULL;
	state.target = NULL;
	state.targetsize = 0;
	state.data = data;

	url_scanner(&state);

	if(newlen) *newlen = state.targetsize;

	return state.target;
}