1 /***************************************************************************
2  *                                  _   _ ____  _
3  *  Project                     ___| | | |  _ \| |
4  *                             / __| | | | |_) | |
5  *                            | (__| |_| |  _ <| |___
6  *                             \___|\___/|_| \_\_____|
7  *
8  * Copyright (C) 1998 - 2015, Daniel Stenberg, <daniel@haxx.se>, et al.
9  *
10  * This software is licensed as described in the file COPYING, which
11  * you should have received as part of this distribution. The terms
12  * are also available at http://curl.haxx.se/docs/copyright.html.
13  *
14  * You may opt to use, copy, modify, merge, publish, distribute and/or sell
15  * copies of the Software, and permit persons to whom the Software is
16  * furnished to do so, under the terms of the COPYING file.
17  *
18  * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
19  * KIND, either express or implied.
20  *
21  ***************************************************************************/
22 
23 /* Escape and unescape URL encoding in strings. The functions return a new
24  * allocated string or NULL if an error occurred.  */
25 
26 #include "curl_setup.h"
27 
28 #include <curl/curl.h>
29 
30 #include "urldata.h"
31 #include "warnless.h"
32 #include "non-ascii.h"
33 #include "escape.h"
34 #include "curl_printf.h"
35 
36 /* The last #include files should be: */
37 #include "curl_memory.h"
38 #include "memdebug.h"
39 
40 /* Portable character check (remember EBCDIC). Do not use isalnum() because
41    its behavior is altered by the current locale.
42    See http://tools.ietf.org/html/rfc3986#section-2.3
43 */
Curl_isunreserved(unsigned char in)44 static bool Curl_isunreserved(unsigned char in)
45 {
46   switch (in) {
47     case '0': case '1': case '2': case '3': case '4':
48     case '5': case '6': case '7': case '8': case '9':
49     case 'a': case 'b': case 'c': case 'd': case 'e':
50     case 'f': case 'g': case 'h': case 'i': case 'j':
51     case 'k': case 'l': case 'm': case 'n': case 'o':
52     case 'p': case 'q': case 'r': case 's': case 't':
53     case 'u': case 'v': case 'w': case 'x': case 'y': case 'z':
54     case 'A': case 'B': case 'C': case 'D': case 'E':
55     case 'F': case 'G': case 'H': case 'I': case 'J':
56     case 'K': case 'L': case 'M': case 'N': case 'O':
57     case 'P': case 'Q': case 'R': case 'S': case 'T':
58     case 'U': case 'V': case 'W': case 'X': case 'Y': case 'Z':
59     case '-': case '.': case '_': case '~':
60       return TRUE;
61     default:
62       break;
63   }
64   return FALSE;
65 }
66 
67 /* for ABI-compatibility with previous versions */
curl_escape(const char * string,int inlength)68 char *curl_escape(const char *string, int inlength)
69 {
70   return curl_easy_escape(NULL, string, inlength);
71 }
72 
73 /* for ABI-compatibility with previous versions */
curl_unescape(const char * string,int length)74 char *curl_unescape(const char *string, int length)
75 {
76   return curl_easy_unescape(NULL, string, length, NULL);
77 }
78 
curl_easy_escape(CURL * handle,const char * string,int inlength)79 char *curl_easy_escape(CURL *handle, const char *string, int inlength)
80 {
81   size_t alloc = (inlength?(size_t)inlength:strlen(string))+1;
82   char *ns;
83   char *testing_ptr = NULL;
84   unsigned char in; /* we need to treat the characters unsigned */
85   size_t newlen = alloc;
86   size_t strindex=0;
87   size_t length;
88   CURLcode result;
89 
90   ns = malloc(alloc);
91   if(!ns)
92     return NULL;
93 
94   length = alloc-1;
95   while(length--) {
96     in = *string;
97 
98     if(Curl_isunreserved(in))
99       /* just copy this */
100       ns[strindex++]=in;
101     else {
102       /* encode it */
103       newlen += 2; /* the size grows with two, since this'll become a %XX */
104       if(newlen > alloc) {
105         alloc *= 2;
106         testing_ptr = realloc(ns, alloc);
107         if(!testing_ptr) {
108           free( ns );
109           return NULL;
110         }
111         else {
112           ns = testing_ptr;
113         }
114       }
115 
116       result = Curl_convert_to_network(handle, &in, 1);
117       if(result) {
118         /* Curl_convert_to_network calls failf if unsuccessful */
119         free(ns);
120         return NULL;
121       }
122 
123       snprintf(&ns[strindex], 4, "%%%02X", in);
124 
125       strindex+=3;
126     }
127     string++;
128   }
129   ns[strindex]=0; /* terminate it */
130   return ns;
131 }
132 
133 /*
134  * Curl_urldecode() URL decodes the given string.
135  *
136  * Optionally detects control characters (byte codes lower than 32) in the
137  * data and rejects such data.
138  *
139  * Returns a pointer to a malloced string in *ostring with length given in
140  * *olen. If length == 0, the length is assumed to be strlen(string).
141  *
142  */
Curl_urldecode(struct SessionHandle * data,const char * string,size_t length,char ** ostring,size_t * olen,bool reject_ctrl)143 CURLcode Curl_urldecode(struct SessionHandle *data,
144                         const char *string, size_t length,
145                         char **ostring, size_t *olen,
146                         bool reject_ctrl)
147 {
148   size_t alloc = (length?length:strlen(string))+1;
149   char *ns = malloc(alloc);
150   unsigned char in;
151   size_t strindex=0;
152   unsigned long hex;
153   CURLcode result;
154 
155   if(!ns)
156     return CURLE_OUT_OF_MEMORY;
157 
158   while(--alloc > 0) {
159     in = *string;
160     if(('%' == in) && (alloc > 2) &&
161        ISXDIGIT(string[1]) && ISXDIGIT(string[2])) {
162       /* this is two hexadecimal digits following a '%' */
163       char hexstr[3];
164       char *ptr;
165       hexstr[0] = string[1];
166       hexstr[1] = string[2];
167       hexstr[2] = 0;
168 
169       hex = strtoul(hexstr, &ptr, 16);
170 
171       in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */
172 
173       result = Curl_convert_from_network(data, &in, 1);
174       if(result) {
175         /* Curl_convert_from_network calls failf if unsuccessful */
176         free(ns);
177         return result;
178       }
179 
180       string+=2;
181       alloc-=2;
182     }
183 
184     if(reject_ctrl && (in < 0x20)) {
185       free(ns);
186       return CURLE_URL_MALFORMAT;
187     }
188 
189     ns[strindex++] = in;
190     string++;
191   }
192   ns[strindex]=0; /* terminate it */
193 
194   if(olen)
195     /* store output size */
196     *olen = strindex;
197 
198   /* store output string */
199   *ostring = ns;
200 
201   return CURLE_OK;
202 }
203 
204 /*
205  * Unescapes the given URL escaped string of given length. Returns a
206  * pointer to a malloced string with length given in *olen.
207  * If length == 0, the length is assumed to be strlen(string).
208  * If olen == NULL, no output length is stored.
209  */
curl_easy_unescape(CURL * handle,const char * string,int length,int * olen)210 char *curl_easy_unescape(CURL *handle, const char *string, int length,
211                          int *olen)
212 {
213   char *str = NULL;
214   size_t inputlen = length;
215   size_t outputlen;
216   CURLcode res = Curl_urldecode(handle, string, inputlen, &str, &outputlen,
217                                 FALSE);
218   if(res)
219     return NULL;
220   if(olen)
221     *olen = curlx_uztosi(outputlen);
222   return str;
223 }
224 
225 /* For operating systems/environments that use different malloc/free
226    systems for the app and for this library, we provide a free that uses
227    the library's memory system */
curl_free(void * p)228 void curl_free(void *p)
229 {
230   free(p);
231 }
232