stream/url.c

   1 /*
   2  * URL Helper
   3  *
   4  * Copyright (C) 2001 Bertrand Baudet <bertrand_baudet@yahoo.com>
   5  *
   6  * This file is part of MPlayer.
   7  *
   8  * MPlayer is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * MPlayer is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with MPlayer; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  21  */
  22
  23 #include <string.h>
  24 #include <stdlib.h>
  25 #include <stdio.h>
  26 #include <ctype.h>
  27 #include <inttypes.h>
  28
  29 #include "url.h"
  30 #include "mp_msg.h"
  31
  32 #ifndef SIZE_MAX
  33 #define SIZE_MAX ((size_t)-1)
  34 #endif
  35
  36 URL_t *url_redirect(URL_t **url, const char *redir) {
  37   URL_t *u = *url;
  38   URL_t *res;
  39   if (!strchr(redir, '/') || *redir == '/') {
  40     char *tmp;
  41     char *newurl = malloc(strlen(u->url) + strlen(redir) + 1);
  42     strcpy(newurl, u->url);
  43     if (*redir == '/') {
  44       redir++;
  45       tmp = strstr(newurl, "://");
  46       if (tmp) tmp = strchr(tmp + 3, '/');
  47     } else
  48       tmp = strrchr(newurl, '/');
  49     if (tmp) tmp[1] = 0;
  50     strcat(newurl, redir);
  51     res = url_new(newurl);
  52     free(newurl);
  53   } else
  54     res = url_new(redir);
  55   url_free(u);
  56   *url = res;
  57   return res;
  58 }
  59
  60 static int make_noauth_url(URL_t *url, char *dst, int dst_size)
  61 {
  62     if (url->port)
  63         return snprintf(dst, dst_size, "%s://%s:%d%s", url->protocol,
  64                         url->hostname, url->port, url->file);
  65     else
  66         return snprintf(dst, dst_size, "%s://%s%s", url->protocol,
  67                         url->hostname, url->file);
  68 }
  69
  70 int make_http_proxy_url(URL_t *proxy, const char *host_url, char *dst,
  71                         int dst_size)
  72 {
  73     if (proxy->username)
  74         return snprintf(dst, dst_size, "http_proxy://%s:%s@%s:%d/%s",
  75                         proxy->username,
  76                         proxy->password ? proxy->password : "",
  77                         proxy->hostname, proxy->port, host_url);
  78     else
  79         return snprintf(dst, dst_size, "http_proxy://%s:%d/%s",
  80                         proxy->hostname, proxy->port, host_url);
  81 }
  82
  83 URL_t*
  84 url_new(const char* url) {
  85         int pos1, pos2,v6addr = 0, noauth_len;
  86         URL_t* Curl = NULL;
  87         char *escfilename=NULL;
  88         char *ptr1=NULL, *ptr2=NULL, *ptr3=NULL, *ptr4=NULL;
  89         int jumpSize = 3;
  90
  91         if( url==NULL ) return NULL;
  92
  93         if (strlen(url) > (SIZE_MAX / 3 - 1)) {
  94                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
  95                 goto err_out;
  96         }
  97         escfilename=malloc(strlen(url)*3+1);
  98         if (!escfilename ) {
  99                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 100                 goto err_out;
 101         }
 102
 103         // Create the URL container
 104         Curl = calloc(1, sizeof(*Curl));
 105         if( Curl==NULL ) {
 106                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 107                 goto err_out;
 108         }
 109
 110         url_escape_string(escfilename,url);
 111
 112         // Copy the url in the URL container
 113         Curl->url = strdup(escfilename);
 114         if( Curl->url==NULL ) {
 115                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 116                 goto err_out;
 117         }
 118         mp_msg(MSGT_OPEN,MSGL_V,"Filename for url is now %s\n",escfilename);
 119
 120         // extract the protocol
 121         ptr1 = strstr(escfilename, "://");
 122         if( ptr1==NULL ) {
 123                 // Check for a special case: "sip:" (without "//"):
 124                 if (strstr(escfilename, "sip:") == escfilename) {
 125                         ptr1 = (char *)&url[3]; // points to ':'
 126                         jumpSize = 1;
 127                 } else {
 128                         mp_msg(MSGT_NETWORK,MSGL_V,"Not an URL!\n");
 129                         goto err_out;
 130                 }
 131         }
 132         pos1 = ptr1-escfilename;
 133         Curl->protocol = malloc(pos1+1);
 134         if( Curl->protocol==NULL ) {
 135                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 136                 goto err_out;
 137         }
 138         strncpy(Curl->protocol, escfilename, pos1);
 139         Curl->protocol[pos1] = '\0';
 140
 141         // jump the "://"
 142         ptr1 += jumpSize;
 143         pos1 += jumpSize;
 144
 145         // check if a username:password is given
 146         ptr2 = strstr(ptr1, "@");
 147         ptr3 = strstr(ptr1, "/");
 148         if( ptr3!=NULL && ptr3<ptr2 ) {
 149                 // it isn't really a username but rather a part of the path
 150                 ptr2 = NULL;
 151         }
 152         if( ptr2!=NULL ) {
 153                 // We got something, at least a username...
 154                 int len = ptr2-ptr1;
 155                 Curl->username = malloc(len+1);
 156                 if( Curl->username==NULL ) {
 157                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 158                         goto err_out;
 159                 }
 160                 strncpy(Curl->username, ptr1, len);
 161                 Curl->username[len] = '\0';
 162
 163                 ptr3 = strstr(ptr1, ":");
 164                 if( ptr3!=NULL && ptr3<ptr2 ) {
 165                         // We also have a password
 166                         int len2 = ptr2-ptr3-1;
 167                         Curl->username[ptr3-ptr1]='\0';
 168                         Curl->password = malloc(len2+1);
 169                         if( Curl->password==NULL ) {
 170                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 171                                 goto err_out;
 172                         }
 173                         strncpy( Curl->password, ptr3+1, len2);
 174                         Curl->password[len2]='\0';
 175                         url_unescape_string(Curl->password, Curl->password);
 176                 }
 177                 url_unescape_string(Curl->username, Curl->username);
 178                 ptr1 = ptr2+1;
 179                 pos1 = ptr1-escfilename;
 180         }
 181
 182         // before looking for a port number check if we have an IPv6 type numeric address
 183         // in IPv6 URL the numeric address should be inside square braces.
 184         ptr2 = strstr(ptr1, "[");
 185         ptr3 = strstr(ptr1, "]");
 186         ptr4 = strstr(ptr1, "/");
 187         if( ptr2!=NULL && ptr3!=NULL && ptr2 < ptr3 && (!ptr4 || ptr4 > ptr3)) {
 188                 // we have an IPv6 numeric address
 189                 ptr1++;
 190                 pos1++;
 191                 ptr2 = ptr3;
 192                 v6addr = 1;
 193         } else {
 194                 ptr2 = ptr1;
 195
 196         }
 197
 198         // look if the port is given
 199         ptr2 = strstr(ptr2, ":");
 200         // If the : is after the first / it isn't the port
 201         ptr3 = strstr(ptr1, "/");
 202         if(ptr3 && ptr3 - ptr2 < 0) ptr2 = NULL;
 203         if( ptr2==NULL ) {
 204                 // No port is given
 205                 // Look if a path is given
 206                 if( ptr3==NULL ) {
 207                         // No path/filename
 208                         // So we have an URL like http://www.hostname.com
 209                         pos2 = strlen(escfilename);
 210                 } else {
 211                         // We have an URL like http://www.hostname.com/file.txt
 212                         pos2 = ptr3-escfilename;
 213                 }
 214         } else {
 215                 // We have an URL beginning like http://www.hostname.com:1212
 216                 // Get the port number
 217                 Curl->port = atoi(ptr2+1);
 218                 pos2 = ptr2-escfilename;
 219         }
 220         if( v6addr ) pos2--;
 221         // copy the hostname in the URL container
 222         Curl->hostname = malloc(pos2-pos1+1);
 223         if( Curl->hostname==NULL ) {
 224                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 225                 goto err_out;
 226         }
 227         strncpy(Curl->hostname, ptr1, pos2-pos1);
 228         Curl->hostname[pos2-pos1] = '\0';
 229
 230         // Look if a path is given
 231         ptr2 = strstr(ptr1, "/");
 232         if( ptr2!=NULL ) {
 233                 // A path/filename is given
 234                 // check if it's not a trailing '/'
 235                 if( strlen(ptr2)>1 ) {
 236                         // copy the path/filename in the URL container
 237                         Curl->file = strdup(ptr2);
 238                         if( Curl->file==NULL ) {
 239                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 240                                 goto err_out;
 241                         }
 242                 }
 243         }
 244         // Check if a filename was given or set, else set it with '/'
 245         if( Curl->file==NULL ) {
 246                 Curl->file = malloc(2);
 247                 if( Curl->file==NULL ) {
 248                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 249                         goto err_out;
 250                 }
 251                 strcpy(Curl->file, "/");
 252         }
 253
 254         noauth_len = make_noauth_url(Curl, NULL, 0);
 255         if (noauth_len > 0) {
 256                 noauth_len++;
 257                 Curl->noauth_url = malloc(noauth_len);
 258                 if (!Curl->noauth_url) {
 259                         mp_msg(MSGT_NETWORK, MSGL_FATAL, "Memory allocation failed.\n");
 260                         goto err_out;
 261                 }
 262                 make_noauth_url(Curl, Curl->noauth_url, noauth_len);
 263         }
 264
 265         free(escfilename);
 266         return Curl;
 267 err_out:
 268         free(escfilename);
 269         if (Curl) url_free(Curl);
 270         return NULL;
 271 }
 272
 273 void
 274 url_free(URL_t* url) {
 275         if(!url) return;
 276         free(url->url);
 277         free(url->protocol);
 278         free(url->hostname);
 279         free(url->file);
 280         free(url->username);
 281         free(url->password);
 282         free(url);
 283 }
 284
 285
 286 /* Replace escape sequences in an URL (or a part of an URL) */
 287 /* works like strcpy(), but without return argument,
 288    except that outbuf == inbuf is allowed */
 289 void
 290 url_unescape_string(char *outbuf, const char *inbuf)
 291 {
 292         unsigned char c,c1,c2;
 293         int i,len=strlen(inbuf);
 294         for (i=0;i<len;i++){
 295                 c = inbuf[i];
 296                 if (c == '%' && i<len-2) { //must have 2 more chars
 297                         c1 = toupper(inbuf[i+1]); // we need uppercase characters
 298                         c2 = toupper(inbuf[i+2]);
 299                         if (    ((c1>='0' && c1<='9') || (c1>='A' && c1<='F')) &&
 300                                 ((c2>='0' && c2<='9') || (c2>='A' && c2<='F')) ) {
 301                                 if (c1>='0' && c1<='9') c1-='0';
 302                                 else c1-='A'-10;
 303                                 if (c2>='0' && c2<='9') c2-='0';
 304                                 else c2-='A'-10;
 305                                 c = (c1<<4) + c2;
 306                                 i=i+2; //only skip next 2 chars if valid esc
 307                         }
 308                 }
 309                 *outbuf++ = c;
 310         }
 311         *outbuf++='\0'; //add nullterm to string
 312 }
 313
 314 static void
 315 url_escape_string_part(char *outbuf, const char *inbuf) {
 316         unsigned char c,c1,c2;
 317         int i,len=strlen(inbuf);
 318
 319         for  (i=0;i<len;i++) {
 320                 c = inbuf[i];
 321                 if ((c=='%') && i<len-2 ) { //need 2 more characters
 322                     c1=toupper(inbuf[i+1]); c2=toupper(inbuf[i+2]); // need uppercase chars
 323                    } else {
 324                     c1=129; c2=129; //not escape chars
 325                    }
 326
 327                 if(     (c >= 'A' && c <= 'Z') ||
 328                         (c >= 'a' && c <= 'z') ||
 329                         (c >= '0' && c <= '9')) {
 330                         *outbuf++ = c;
 331                 } else if ( c=='%' && ((c1 >= '0' && c1 <= '9') || (c1 >= 'A' && c1 <= 'F')) &&
 332                            ((c2 >= '0' && c2 <= '9') || (c2 >= 'A' && c2 <= 'F'))) {
 333                                                               // check if part of an escape sequence
 334                             *outbuf++=c;                      // already
 335
 336                                                               // dont escape again
 337                             mp_tmsg(MSGT_NETWORK,MSGL_ERR,"String appears to be already escaped in url_escape %c%c1%c2\n",c,c1,c2);
 338                                                               // error as this should not happen against RFC 2396
 339                                                               // to escape a string twice
 340                 } else {
 341                         /* all others will be escaped */
 342                         c1 = ((c & 0xf0) >> 4);
 343                         c2 = (c & 0x0f);
 344                         if (c1 < 10) c1+='0';
 345                         else c1+='A'-10;
 346                         if (c2 < 10) c2+='0';
 347                         else c2+='A'-10;
 348                         *outbuf++ = '%';
 349                         *outbuf++ = c1;
 350                         *outbuf++ = c2;
 351                 }
 352         }
 353         *outbuf++='\0';
 354 }
 355
 356 /* Replace specific characters in the URL string by an escape sequence */
 357 /* works like strcpy(), but without return argument */
 358 void
 359 url_escape_string(char *outbuf, const char *inbuf) {
 360         unsigned char c;
 361         int i = 0,j,len = strlen(inbuf);
 362         char* tmp,*unesc = NULL, *in;
 363
 364         // Look if we have an ip6 address, if so skip it there is
 365         // no need to escape anything in there.
 366         tmp = strstr(inbuf,"://[");
 367         if(tmp) {
 368                 tmp = strchr(tmp+4,']');
 369                 if(tmp && (tmp[1] == '/' || tmp[1] == ':' ||
 370                            tmp[1] == '\0')) {
 371                         i = tmp+1-inbuf;
 372                         strncpy(outbuf,inbuf,i);
 373                         outbuf += i;
 374                         tmp = NULL;
 375                 }
 376         }
 377
 378         tmp = NULL;
 379         while(i < len) {
 380                 // look for the next char that must be kept
 381                 for  (j=i;j<len;j++) {
 382                         c = inbuf[j];
 383                         if(c=='-' || c=='_' || c=='.' || c=='!' || c=='~' ||    /* mark characters */
 384                            c=='*' || c=='\'' || c=='(' || c==')' ||             /* do not touch escape character */
 385                            c==';' || c=='/' || c=='?' || c==':' || c=='@' ||    /* reserved characters */
 386                            c=='&' || c=='=' || c=='+' || c=='$' || c==',')      /* see RFC 2396 */
 387                                 break;
 388                 }
 389                 // we are on a reserved char, write it out
 390                 if(j == i) {
 391                         *outbuf++ = c;
 392                         i++;
 393                         continue;
 394                 }
 395                 // we found one, take that part of the string
 396                 if(j < len) {
 397                         if(!tmp) tmp = malloc(len+1);
 398                         strncpy(tmp,inbuf+i,j-i);
 399                         tmp[j-i] = '\0';
 400                         in = tmp;
 401                 } else // take the rest of the string
 402                         in = (char*)inbuf+i;
 403
 404                 if(!unesc) unesc = malloc(len+1);
 405                 // unescape first to avoid escaping escape
 406                 url_unescape_string(unesc,in);
 407                 // then escape, including mark and other reserved chars
 408                 // that can come from escape sequences
 409                 url_escape_string_part(outbuf,unesc);
 410                 outbuf += strlen(outbuf);
 411                 i += strlen(in);
 412         }
 413         *outbuf = '\0';
 414         free(tmp);
 415         free(unesc);
 416 }
 417
 418 #ifdef URL_DEBUG
 419 void
 420 url_debug(const URL_t *url) {
 421         if( url==NULL ) {
 422                 printf("URL pointer NULL\n");
 423                 return;
 424         }
 425         if( url->url!=NULL ) {
 426                 printf("url=%s\n", url->url );
 427         }
 428         if( url->protocol!=NULL ) {
 429                 printf("protocol=%s\n", url->protocol );
 430         }
 431         if( url->hostname!=NULL ) {
 432                 printf("hostname=%s\n", url->hostname );
 433         }
 434         printf("port=%d\n", url->port );
 435         if( url->file!=NULL ) {
 436                 printf("file=%s\n", url->file );
 437         }
 438         if( url->username!=NULL ) {
 439                 printf("username=%s\n", url->username );
 440         }
 441         if( url->password!=NULL ) {
 442                 printf("password=%s\n", url->password );
 443         }
 444 }
 445 #endif /* URL_DEBUG */