File:  [ELWIX - Embedded LightWeight unIX -] / embedaddon / php / ext / standard / url.c
Revision 1.1.1.5 (vendor branch): download - view: text, annotated - select for diffs - revision graph
Sun Jun 15 20:03:57 2014 UTC (10 years, 1 month ago) by misho
Branches: php, MAIN
CVS tags: v5_4_29, HEAD
php 5.4.29

    1: /*
    2:    +----------------------------------------------------------------------+
    3:    | PHP Version 5                                                        |
    4:    +----------------------------------------------------------------------+
    5:    | Copyright (c) 1997-2014 The PHP Group                                |
    6:    +----------------------------------------------------------------------+
    7:    | This source file is subject to version 3.01 of the PHP license,      |
    8:    | that is bundled with this package in the file LICENSE, and is        |
    9:    | available through the world-wide-web at the following url:           |
   10:    | http://www.php.net/license/3_01.txt                                  |
   11:    | If you did not receive a copy of the PHP license and are unable to   |
   12:    | obtain it through the world-wide-web, please send a note to          |
   13:    | license@php.net so we can mail you a copy immediately.               |
   14:    +----------------------------------------------------------------------+
   15:    | Author: Jim Winstead <jimw@php.net>                                  |
   16:    +----------------------------------------------------------------------+
   17:  */
   18: /* $Id: url.c,v 1.1.1.5 2014/06/15 20:03:57 misho Exp $ */
   19: 
   20: #include <stdlib.h>
   21: #include <string.h>
   22: #include <ctype.h>
   23: #include <sys/types.h>
   24: 
   25: #include "php.h"
   26: 
   27: #include "url.h"
   28: #include "file.h"
   29: #ifdef _OSD_POSIX
   30: #ifndef APACHE
   31: #error On this EBCDIC platform, PHP is only supported as an Apache module.
   32: #else /*APACHE*/
   33: #ifndef CHARSET_EBCDIC
   34: #define CHARSET_EBCDIC /* this machine uses EBCDIC, not ASCII! */
   35: #endif
   36: #include "ebcdic.h"
   37: #endif /*APACHE*/
   38: #endif /*_OSD_POSIX*/
   39: 
   40: /* {{{ free_url
   41:  */
   42: PHPAPI void php_url_free(php_url *theurl)
   43: {
   44: 	if (theurl->scheme)
   45: 		efree(theurl->scheme);
   46: 	if (theurl->user)
   47: 		efree(theurl->user);
   48: 	if (theurl->pass)
   49: 		efree(theurl->pass);
   50: 	if (theurl->host)
   51: 		efree(theurl->host);
   52: 	if (theurl->path)
   53: 		efree(theurl->path);
   54: 	if (theurl->query)
   55: 		efree(theurl->query);
   56: 	if (theurl->fragment)
   57: 		efree(theurl->fragment);
   58: 	efree(theurl);
   59: }
   60: /* }}} */
   61: 
   62: /* {{{ php_replace_controlchars
   63:  */
   64: PHPAPI char *php_replace_controlchars_ex(char *str, int len)
   65: {
   66: 	unsigned char *s = (unsigned char *)str;
   67: 	unsigned char *e = (unsigned char *)str + len;
   68: 	
   69: 	if (!str) {
   70: 		return (NULL);
   71: 	}
   72: 	
   73: 	while (s < e) {
   74: 	    
   75: 		if (iscntrl(*s)) {
   76: 			*s='_';
   77: 		}	
   78: 		s++;
   79: 	}
   80: 	
   81: 	return (str);
   82: } 
   83: /* }}} */
   84: 
   85: PHPAPI char *php_replace_controlchars(char *str)
   86: {
   87: 	return php_replace_controlchars_ex(str, strlen(str));
   88: } 
   89: 
   90: PHPAPI php_url *php_url_parse(char const *str)
   91: {
   92: 	return php_url_parse_ex(str, strlen(str));
   93: }
   94: 
   95: /* {{{ php_url_parse
   96:  */
   97: PHPAPI php_url *php_url_parse_ex(char const *str, int length)
   98: {
   99: 	char port_buf[6];
  100: 	php_url *ret = ecalloc(1, sizeof(php_url));
  101: 	char const *s, *e, *p, *pp, *ue;
  102: 		
  103: 	s = str;
  104: 	ue = s + length;
  105: 
  106: 	/* parse scheme */
  107: 	if ((e = memchr(s, ':', length)) && (e - s)) {
  108: 		/* validate scheme */
  109: 		p = s;
  110: 		while (p < e) {
  111: 			/* scheme = 1*[ lowalpha | digit | "+" | "-" | "." ] */
  112: 			if (!isalpha(*p) && !isdigit(*p) && *p != '+' && *p != '.' && *p != '-') {
  113: 				if (e + 1 < ue) {
  114: 					goto parse_port;
  115: 				} else {
  116: 					goto just_path;
  117: 				}
  118: 			}
  119: 			p++;
  120: 		}
  121: 	
  122: 		if (*(e + 1) == '\0') { /* only scheme is available */
  123: 			ret->scheme = estrndup(s, (e - s));
  124: 			php_replace_controlchars_ex(ret->scheme, (e - s));
  125: 			goto end;
  126: 		}
  127: 
  128: 		/* 
  129: 		 * certain schemas like mailto: and zlib: may not have any / after them
  130: 		 * this check ensures we support those.
  131: 		 */
  132: 		if (*(e+1) != '/') {
  133: 			/* check if the data we get is a port this allows us to 
  134: 			 * correctly parse things like a.com:80
  135: 			 */
  136: 			p = e + 1;
  137: 			while (isdigit(*p)) {
  138: 				p++;
  139: 			}
  140: 			
  141: 			if ((*p == '\0' || *p == '/') && (p - e) < 7) {
  142: 				goto parse_port;
  143: 			}
  144: 			
  145: 			ret->scheme = estrndup(s, (e-s));
  146: 			php_replace_controlchars_ex(ret->scheme, (e - s));
  147: 			
  148: 			length -= ++e - s;
  149: 			s = e;
  150: 			goto just_path;
  151: 		} else {
  152: 			ret->scheme = estrndup(s, (e-s));
  153: 			php_replace_controlchars_ex(ret->scheme, (e - s));
  154: 		
  155: 			if (*(e+2) == '/') {
  156: 				s = e + 3;
  157: 				if (!strncasecmp("file", ret->scheme, sizeof("file"))) {
  158: 					if (*(e + 3) == '/') {
  159: 						/* support windows drive letters as in:
  160: 						   file:///c:/somedir/file.txt
  161: 						*/
  162: 						if (*(e + 5) == ':') {
  163: 							s = e + 4;
  164: 						}
  165: 						goto nohost;
  166: 					}
  167: 				}
  168: 			} else {
  169: 				if (!strncasecmp("file", ret->scheme, sizeof("file"))) {
  170: 					s = e + 1;
  171: 					goto nohost;
  172: 				} else {
  173: 					length -= ++e - s;
  174: 					s = e;
  175: 					goto just_path;
  176: 				}	
  177: 			}
  178: 		}	
  179: 	} else if (e) { /* no scheme; starts with colon: look for port */
  180: 		parse_port:
  181: 		p = e + 1;
  182: 		pp = p;
  183: 
  184: 		while (pp-p < 6 && isdigit(*pp)) {
  185: 			pp++;
  186: 		}
  187: 
  188: 		if (pp - p > 0 && pp - p < 6 && (*pp == '/' || *pp == '\0')) {
  189: 			long port;
  190: 			memcpy(port_buf, p, (pp - p));
  191: 			port_buf[pp - p] = '\0';
  192: 			port = strtol(port_buf, NULL, 10);
  193: 			if (port > 0 && port <= 65535) {
  194: 				ret->port = (unsigned short) port;
  195: 			} else {
  196: 				STR_FREE(ret->scheme);
  197: 				efree(ret);
  198: 				return NULL;
  199: 			}
  200: 		} else if (p == pp && *pp == '\0') {
  201: 			STR_FREE(ret->scheme);
  202: 			efree(ret);
  203: 			return NULL;
  204: 		} else if (*s == '/' && *(s+1) == '/') { /* relative-scheme URL */
  205: 			s += 2;
  206: 		} else {
  207: 			goto just_path;
  208: 		}
  209: 	} else if (*s == '/' && *(s+1) == '/') { /* relative-scheme URL */
  210: 		s += 2;
  211: 	} else {
  212: 		just_path:
  213: 		ue = s + length;
  214: 		goto nohost;
  215: 	}
  216: 	
  217: 	e = ue;
  218: 	
  219: 	if (!(p = memchr(s, '/', (ue - s)))) {
  220: 		char *query, *fragment;
  221: 
  222: 		query = memchr(s, '?', (ue - s));
  223: 		fragment = memchr(s, '#', (ue - s));
  224: 
  225: 		if (query && fragment) {
  226: 			if (query > fragment) {
  227: 				e = fragment;
  228: 			} else {
  229: 				e = query;
  230: 			}
  231: 		} else if (query) {
  232: 			e = query;
  233: 		} else if (fragment) {
  234: 			e = fragment;
  235: 		}
  236: 	} else {
  237: 		e = p;
  238: 	}	
  239: 		
  240: 	/* check for login and password */
  241: 	if ((p = zend_memrchr(s, '@', (e-s)))) {
  242: 		if ((pp = memchr(s, ':', (p-s)))) {
  243: 			if ((pp-s) > 0) {
  244: 				ret->user = estrndup(s, (pp-s));
  245: 				php_replace_controlchars_ex(ret->user, (pp - s));
  246: 			}	
  247: 		
  248: 			pp++;
  249: 			if (p-pp > 0) {
  250: 				ret->pass = estrndup(pp, (p-pp));
  251: 				php_replace_controlchars_ex(ret->pass, (p-pp));
  252: 			}	
  253: 		} else {
  254: 			ret->user = estrndup(s, (p-s));
  255: 			php_replace_controlchars_ex(ret->user, (p-s));
  256: 		}
  257: 		
  258: 		s = p + 1;
  259: 	}
  260: 
  261: 	/* check for port */
  262: 	if (*s == '[' && *(e-1) == ']') {
  263: 		/* Short circuit portscan, 
  264: 		   we're dealing with an 
  265: 		   IPv6 embedded address */
  266: 		p = s;
  267: 	} else {
  268: 		/* memrchr is a GNU specific extension
  269: 		   Emulate for wide compatability */
  270: 		for(p = e; p >= s && *p != ':'; p--);
  271: 	}
  272: 
  273: 	if (p >= s && *p == ':') {
  274: 		if (!ret->port) {
  275: 			p++;
  276: 			if (e-p > 5) { /* port cannot be longer then 5 characters */
  277: 				STR_FREE(ret->scheme);
  278: 				STR_FREE(ret->user);
  279: 				STR_FREE(ret->pass);
  280: 				efree(ret);
  281: 				return NULL;
  282: 			} else if (e - p > 0) {
  283: 				long port;
  284: 				memcpy(port_buf, p, (e - p));
  285: 				port_buf[e - p] = '\0';
  286: 				port = strtol(port_buf, NULL, 10);
  287: 				if (port > 0 && port <= 65535) {
  288: 					ret->port = (unsigned short)port;
  289: 				} else {
  290: 					STR_FREE(ret->scheme);
  291: 					STR_FREE(ret->user);
  292: 					STR_FREE(ret->pass);
  293: 					efree(ret);
  294: 					return NULL;
  295: 				}
  296: 			}
  297: 			p--;
  298: 		}	
  299: 	} else {
  300: 		p = e;
  301: 	}
  302: 	
  303: 	/* check if we have a valid host, if we don't reject the string as url */
  304: 	if ((p-s) < 1) {
  305: 		STR_FREE(ret->scheme);
  306: 		STR_FREE(ret->user);
  307: 		STR_FREE(ret->pass);
  308: 		efree(ret);
  309: 		return NULL;
  310: 	}
  311: 
  312: 	ret->host = estrndup(s, (p-s));
  313: 	php_replace_controlchars_ex(ret->host, (p - s));
  314: 	
  315: 	if (e == ue) {
  316: 		return ret;
  317: 	}
  318: 	
  319: 	s = e;
  320: 	
  321: 	nohost:
  322: 	
  323: 	if ((p = memchr(s, '?', (ue - s)))) {
  324: 		pp = strchr(s, '#');
  325: 
  326: 		if (pp && pp < p) {
  327: 			if (pp - s) {
  328: 				ret->path = estrndup(s, (pp-s));
  329: 				php_replace_controlchars_ex(ret->path, (pp - s));
  330: 			}
  331: 			p = pp;
  332: 			goto label_parse;
  333: 		}
  334: 	
  335: 		if (p - s) {
  336: 			ret->path = estrndup(s, (p-s));
  337: 			php_replace_controlchars_ex(ret->path, (p - s));
  338: 		}	
  339: 	
  340: 		if (pp) {
  341: 			if (pp - ++p) { 
  342: 				ret->query = estrndup(p, (pp-p));
  343: 				php_replace_controlchars_ex(ret->query, (pp - p));
  344: 			}
  345: 			p = pp;
  346: 			goto label_parse;
  347: 		} else if (++p - ue) {
  348: 			ret->query = estrndup(p, (ue-p));
  349: 			php_replace_controlchars_ex(ret->query, (ue - p));
  350: 		}
  351: 	} else if ((p = memchr(s, '#', (ue - s)))) {
  352: 		if (p - s) {
  353: 			ret->path = estrndup(s, (p-s));
  354: 			php_replace_controlchars_ex(ret->path, (p - s));
  355: 		}	
  356: 		
  357: 		label_parse:
  358: 		p++;
  359: 		
  360: 		if (ue - p) {
  361: 			ret->fragment = estrndup(p, (ue-p));
  362: 			php_replace_controlchars_ex(ret->fragment, (ue - p));
  363: 		}	
  364: 	} else {
  365: 		ret->path = estrndup(s, (ue-s));
  366: 		php_replace_controlchars_ex(ret->path, (ue - s));
  367: 	}
  368: end:
  369: 	return ret;
  370: }
  371: /* }}} */
  372: 
  373: /* {{{ proto mixed parse_url(string url, [int url_component])
  374:    Parse a URL and return its components */
  375: PHP_FUNCTION(parse_url)
  376: {
  377: 	char *str;
  378: 	int str_len;
  379: 	php_url *resource;
  380: 	long key = -1;
  381: 
  382: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|l", &str, &str_len, &key) == FAILURE) {
  383: 		return;
  384: 	}
  385: 
  386: 	resource = php_url_parse_ex(str, str_len);
  387: 	if (resource == NULL) {
  388: 		/* @todo Find a method to determine why php_url_parse_ex() failed */
  389: 		RETURN_FALSE;
  390: 	}
  391: 
  392: 	if (key > -1) {
  393: 		switch (key) {
  394: 			case PHP_URL_SCHEME:
  395: 				if (resource->scheme != NULL) RETVAL_STRING(resource->scheme, 1);
  396: 				break;
  397: 			case PHP_URL_HOST:
  398: 				if (resource->host != NULL) RETVAL_STRING(resource->host, 1);
  399: 				break;
  400: 			case PHP_URL_PORT:
  401: 				if (resource->port != 0) RETVAL_LONG(resource->port);
  402: 				break;
  403: 			case PHP_URL_USER:
  404: 				if (resource->user != NULL) RETVAL_STRING(resource->user, 1);
  405: 				break;
  406: 			case PHP_URL_PASS:
  407: 				if (resource->pass != NULL) RETVAL_STRING(resource->pass, 1);
  408: 				break;
  409: 			case PHP_URL_PATH:
  410: 				if (resource->path != NULL) RETVAL_STRING(resource->path, 1);
  411: 				break;
  412: 			case PHP_URL_QUERY:
  413: 				if (resource->query != NULL) RETVAL_STRING(resource->query, 1);
  414: 				break;
  415: 			case PHP_URL_FRAGMENT:
  416: 				if (resource->fragment != NULL) RETVAL_STRING(resource->fragment, 1);
  417: 				break;
  418: 			default:
  419: 				php_error_docref(NULL TSRMLS_CC, E_WARNING, "Invalid URL component identifier %ld", key);
  420: 				RETVAL_FALSE;
  421: 		}
  422: 		goto done;
  423: 	}
  424: 
  425: 	/* allocate an array for return */
  426: 	array_init(return_value);
  427: 
  428:     /* add the various elements to the array */
  429: 	if (resource->scheme != NULL)
  430: 		add_assoc_string(return_value, "scheme", resource->scheme, 1);
  431: 	if (resource->host != NULL)
  432: 		add_assoc_string(return_value, "host", resource->host, 1);
  433: 	if (resource->port != 0)
  434: 		add_assoc_long(return_value, "port", resource->port);
  435: 	if (resource->user != NULL)
  436: 		add_assoc_string(return_value, "user", resource->user, 1);
  437: 	if (resource->pass != NULL)
  438: 		add_assoc_string(return_value, "pass", resource->pass, 1);
  439: 	if (resource->path != NULL)
  440: 		add_assoc_string(return_value, "path", resource->path, 1);
  441: 	if (resource->query != NULL)
  442: 		add_assoc_string(return_value, "query", resource->query, 1);
  443: 	if (resource->fragment != NULL)
  444: 		add_assoc_string(return_value, "fragment", resource->fragment, 1);
  445: done:	
  446: 	php_url_free(resource);
  447: }
  448: /* }}} */
  449: 
  450: /* {{{ php_htoi
  451:  */
  452: static int php_htoi(char *s)
  453: {
  454: 	int value;
  455: 	int c;
  456: 
  457: 	c = ((unsigned char *)s)[0];
  458: 	if (isupper(c))
  459: 		c = tolower(c);
  460: 	value = (c >= '0' && c <= '9' ? c - '0' : c - 'a' + 10) * 16;
  461: 
  462: 	c = ((unsigned char *)s)[1];
  463: 	if (isupper(c))
  464: 		c = tolower(c);
  465: 	value += c >= '0' && c <= '9' ? c - '0' : c - 'a' + 10;
  466: 
  467: 	return (value);
  468: }
  469: /* }}} */
  470: 
  471: /* rfc1738:
  472: 
  473:    ...The characters ";",
  474:    "/", "?", ":", "@", "=" and "&" are the characters which may be
  475:    reserved for special meaning within a scheme...
  476: 
  477:    ...Thus, only alphanumerics, the special characters "$-_.+!*'(),", and
  478:    reserved characters used for their reserved purposes may be used
  479:    unencoded within a URL...
  480: 
  481:    For added safety, we only leave -_. unencoded.
  482:  */
  483: 
  484: static unsigned char hexchars[] = "0123456789ABCDEF";
  485: 
  486: /* {{{ php_url_encode
  487:  */
  488: PHPAPI char *php_url_encode(char const *s, int len, int *new_length)
  489: {
  490: 	register unsigned char c;
  491: 	unsigned char *to, *start;
  492: 	unsigned char const *from, *end;
  493: 	
  494: 	from = (unsigned char *)s;
  495: 	end = (unsigned char *)s + len;
  496: 	start = to = (unsigned char *) safe_emalloc(3, len, 1);
  497: 
  498: 	while (from < end) {
  499: 		c = *from++;
  500: 
  501: 		if (c == ' ') {
  502: 			*to++ = '+';
  503: #ifndef CHARSET_EBCDIC
  504: 		} else if ((c < '0' && c != '-' && c != '.') ||
  505: 				   (c < 'A' && c > '9') ||
  506: 				   (c > 'Z' && c < 'a' && c != '_') ||
  507: 				   (c > 'z')) {
  508: 			to[0] = '%';
  509: 			to[1] = hexchars[c >> 4];
  510: 			to[2] = hexchars[c & 15];
  511: 			to += 3;
  512: #else /*CHARSET_EBCDIC*/
  513: 		} else if (!isalnum(c) && strchr("_-.", c) == NULL) {
  514: 			/* Allow only alphanumeric chars and '_', '-', '.'; escape the rest */
  515: 			to[0] = '%';
  516: 			to[1] = hexchars[os_toascii[c] >> 4];
  517: 			to[2] = hexchars[os_toascii[c] & 15];
  518: 			to += 3;
  519: #endif /*CHARSET_EBCDIC*/
  520: 		} else {
  521: 			*to++ = c;
  522: 		}
  523: 	}
  524: 	*to = 0;
  525: 	if (new_length) {
  526: 		*new_length = to - start;
  527: 	}
  528: 	return (char *) start;
  529: }
  530: /* }}} */
  531: 
  532: /* {{{ proto string urlencode(string str)
  533:    URL-encodes string */
  534: PHP_FUNCTION(urlencode)
  535: {
  536: 	char *in_str, *out_str;
  537: 	int in_str_len, out_str_len;
  538: 
  539: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &in_str,
  540: 							  &in_str_len) == FAILURE) {
  541: 		return;
  542: 	}
  543: 
  544: 	out_str = php_url_encode(in_str, in_str_len, &out_str_len);
  545: 	RETURN_STRINGL(out_str, out_str_len, 0);
  546: }
  547: /* }}} */
  548: 
  549: /* {{{ proto string urldecode(string str)
  550:    Decodes URL-encoded string */
  551: PHP_FUNCTION(urldecode)
  552: {
  553: 	char *in_str, *out_str;
  554: 	int in_str_len, out_str_len;
  555: 
  556: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &in_str,
  557: 							  &in_str_len) == FAILURE) {
  558: 		return;
  559: 	}
  560: 
  561: 	out_str = estrndup(in_str, in_str_len);
  562: 	out_str_len = php_url_decode(out_str, in_str_len);
  563: 
  564:     RETURN_STRINGL(out_str, out_str_len, 0);
  565: }
  566: /* }}} */
  567: 
  568: /* {{{ php_url_decode
  569:  */
  570: PHPAPI int php_url_decode(char *str, int len)
  571: {
  572: 	char *dest = str;
  573: 	char *data = str;
  574: 
  575: 	while (len--) {
  576: 		if (*data == '+') {
  577: 			*dest = ' ';
  578: 		}
  579: 		else if (*data == '%' && len >= 2 && isxdigit((int) *(data + 1)) 
  580: 				 && isxdigit((int) *(data + 2))) {
  581: #ifndef CHARSET_EBCDIC
  582: 			*dest = (char) php_htoi(data + 1);
  583: #else
  584: 			*dest = os_toebcdic[(char) php_htoi(data + 1)];
  585: #endif
  586: 			data += 2;
  587: 			len -= 2;
  588: 		} else {
  589: 			*dest = *data;
  590: 		}
  591: 		data++;
  592: 		dest++;
  593: 	}
  594: 	*dest = '\0';
  595: 	return dest - str;
  596: }
  597: /* }}} */
  598: 
  599: /* {{{ php_raw_url_encode
  600:  */
  601: PHPAPI char *php_raw_url_encode(char const *s, int len, int *new_length)
  602: {
  603: 	register int x, y;
  604: 	unsigned char *str;
  605: 
  606: 	str = (unsigned char *) safe_emalloc(3, len, 1);
  607: 	for (x = 0, y = 0; len--; x++, y++) {
  608: 		str[y] = (unsigned char) s[x];
  609: #ifndef CHARSET_EBCDIC
  610: 		if ((str[y] < '0' && str[y] != '-' && str[y] != '.') ||
  611: 			(str[y] < 'A' && str[y] > '9') ||
  612: 			(str[y] > 'Z' && str[y] < 'a' && str[y] != '_') ||
  613: 			(str[y] > 'z' && str[y] != '~')) {
  614: 			str[y++] = '%';
  615: 			str[y++] = hexchars[(unsigned char) s[x] >> 4];
  616: 			str[y] = hexchars[(unsigned char) s[x] & 15];
  617: #else /*CHARSET_EBCDIC*/
  618: 		if (!isalnum(str[y]) && strchr("_-.~", str[y]) != NULL) {
  619: 			str[y++] = '%';
  620: 			str[y++] = hexchars[os_toascii[(unsigned char) s[x]] >> 4];
  621: 			str[y] = hexchars[os_toascii[(unsigned char) s[x]] & 15];
  622: #endif /*CHARSET_EBCDIC*/
  623: 		}
  624: 	}
  625: 	str[y] = '\0';
  626: 	if (new_length) {
  627: 		*new_length = y;
  628: 	}
  629: 	return ((char *) str);
  630: }
  631: /* }}} */
  632: 
  633: /* {{{ proto string rawurlencode(string str)
  634:    URL-encodes string */
  635: PHP_FUNCTION(rawurlencode)
  636: {
  637: 	char *in_str, *out_str;
  638: 	int in_str_len, out_str_len;
  639: 
  640: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &in_str,
  641: 							  &in_str_len) == FAILURE) {
  642: 		return;
  643: 	}
  644: 
  645: 	out_str = php_raw_url_encode(in_str, in_str_len, &out_str_len);
  646: 	RETURN_STRINGL(out_str, out_str_len, 0);
  647: }
  648: /* }}} */
  649: 
  650: /* {{{ proto string rawurldecode(string str)
  651:    Decodes URL-encodes string */
  652: PHP_FUNCTION(rawurldecode)
  653: {
  654: 	char *in_str, *out_str;
  655: 	int in_str_len, out_str_len;
  656: 
  657: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &in_str,
  658: 							  &in_str_len) == FAILURE) {
  659: 		return;
  660: 	}
  661: 
  662: 	out_str = estrndup(in_str, in_str_len);
  663: 	out_str_len = php_raw_url_decode(out_str, in_str_len);
  664: 
  665:     RETURN_STRINGL(out_str, out_str_len, 0);
  666: }
  667: /* }}} */
  668: 
  669: /* {{{ php_raw_url_decode
  670:  */
  671: PHPAPI int php_raw_url_decode(char *str, int len)
  672: {
  673: 	char *dest = str;
  674: 	char *data = str;
  675: 
  676: 	while (len--) {
  677: 		if (*data == '%' && len >= 2 && isxdigit((int) *(data + 1)) 
  678: 			&& isxdigit((int) *(data + 2))) {
  679: #ifndef CHARSET_EBCDIC
  680: 			*dest = (char) php_htoi(data + 1);
  681: #else
  682: 			*dest = os_toebcdic[(char) php_htoi(data + 1)];
  683: #endif
  684: 			data += 2;
  685: 			len -= 2;
  686: 		} else {
  687: 			*dest = *data;
  688: 		}
  689: 		data++;
  690: 		dest++;
  691: 	}
  692: 	*dest = '\0';
  693: 	return dest - str;
  694: }
  695: /* }}} */
  696: 
  697: /* {{{ proto array get_headers(string url[, int format])
  698:    fetches all the headers sent by the server in response to a HTTP request */
  699: PHP_FUNCTION(get_headers)
  700: {
  701: 	char *url;
  702: 	int url_len;
  703: 	php_stream_context *context;
  704: 	php_stream *stream;
  705: 	zval **prev_val, **hdr = NULL, **h;
  706: 	HashPosition pos;
  707: 	HashTable *hashT;
  708: 	long format = 0;
  709:                 
  710: 	if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s|l", &url, &url_len, &format) == FAILURE) {
  711: 		return;
  712: 	}
  713: 	context = FG(default_context) ? FG(default_context) : (FG(default_context) = php_stream_context_alloc(TSRMLS_C));
  714: 
  715: 	if (!(stream = php_stream_open_wrapper_ex(url, "r", REPORT_ERRORS | STREAM_USE_URL | STREAM_ONLY_GET_HEADERS, NULL, context))) {
  716: 		RETURN_FALSE;
  717: 	}
  718: 
  719: 	if (!stream->wrapperdata || Z_TYPE_P(stream->wrapperdata) != IS_ARRAY) {
  720: 		php_stream_close(stream);
  721: 		RETURN_FALSE;
  722: 	}
  723: 
  724: 	array_init(return_value);
  725: 
  726: 	/* check for curl-wrappers that provide headers via a special "headers" element */
  727: 	if (zend_hash_find(HASH_OF(stream->wrapperdata), "headers", sizeof("headers"), (void **)&h) != FAILURE && Z_TYPE_PP(h) == IS_ARRAY) {
  728: 		/* curl-wrappers don't load data until the 1st read */ 
  729: 		if (!Z_ARRVAL_PP(h)->nNumOfElements) {
  730: 			php_stream_getc(stream);
  731: 		}
  732: 		zend_hash_find(HASH_OF(stream->wrapperdata), "headers", sizeof("headers"), (void **)&h);
  733: 		hashT = Z_ARRVAL_PP(h);	
  734: 	} else {
  735: 		hashT = HASH_OF(stream->wrapperdata);
  736: 	}
  737: 
  738: 	zend_hash_internal_pointer_reset_ex(hashT, &pos);
  739: 	while (zend_hash_get_current_data_ex(hashT, (void**)&hdr, &pos) != FAILURE) {
  740: 		if (!hdr || Z_TYPE_PP(hdr) != IS_STRING) {
  741: 			zend_hash_move_forward_ex(hashT, &pos);
  742: 			continue;
  743: 		}
  744: 		if (!format) {
  745: no_name_header:
  746: 			add_next_index_stringl(return_value, Z_STRVAL_PP(hdr), Z_STRLEN_PP(hdr), 1);
  747: 		} else {
  748: 			char c;
  749: 			char *s, *p;
  750: 
  751: 			if ((p = strchr(Z_STRVAL_PP(hdr), ':'))) {
  752: 				c = *p;
  753: 				*p = '\0';
  754: 				s = p + 1;
  755: 				while (isspace((int)*(unsigned char *)s)) {
  756: 					s++;
  757: 				}
  758: 
  759: 				if (zend_hash_find(HASH_OF(return_value), Z_STRVAL_PP(hdr), (p - Z_STRVAL_PP(hdr) + 1), (void **) &prev_val) == FAILURE) {
  760: 					add_assoc_stringl_ex(return_value, Z_STRVAL_PP(hdr), (p - Z_STRVAL_PP(hdr) + 1), s, (Z_STRLEN_PP(hdr) - (s - Z_STRVAL_PP(hdr))), 1);
  761: 				} else { /* some headers may occur more then once, therefor we need to remake the string into an array */
  762: 					convert_to_array(*prev_val);
  763: 					add_next_index_stringl(*prev_val, s, (Z_STRLEN_PP(hdr) - (s - Z_STRVAL_PP(hdr))), 1);
  764: 				}
  765: 
  766: 				*p = c;
  767: 			} else {
  768: 				goto no_name_header;
  769: 			}
  770: 		}
  771: 		zend_hash_move_forward_ex(hashT, &pos);
  772: 	}
  773: 
  774: 	php_stream_close(stream);
  775: }
  776: /* }}} */
  777: 
  778: /*
  779:  * Local variables:
  780:  * tab-width: 4
  781:  * c-basic-offset: 4
  782:  * End:
  783:  * vim600: sw=4 ts=4 fdm=marker
  784:  * vim<600: sw=4 ts=4
  785:  */

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>