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 = malloc(sizeof(URL_t));
 105         if( Curl==NULL ) {
 106                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 107                 goto err_out;
 108         }
 109
 110         // Initialisation of the URL container members
 111         memset( Curl, 0, sizeof(URL_t) );
 112
 113         url_escape_string(escfilename,url);
 114
 115         // Copy the url in the URL container
 116         Curl->url = strdup(escfilename);
 117         if( Curl->url==NULL ) {
 118                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 119                 goto err_out;
 120         }
 121         mp_msg(MSGT_OPEN,MSGL_V,"Filename for url is now %s\n",escfilename);
 122
 123         // extract the protocol
 124         ptr1 = strstr(escfilename, "://");
 125         if( ptr1==NULL ) {
 126                 // Check for a special case: "sip:" (without "//"):
 127                 if (strstr(escfilename, "sip:") == escfilename) {
 128                         ptr1 = (char *)&url[3]; // points to ':'
 129                         jumpSize = 1;
 130                 } else {
 131                         mp_msg(MSGT_NETWORK,MSGL_V,"Not an URL!\n");
 132                         goto err_out;
 133                 }
 134         }
 135         pos1 = ptr1-escfilename;
 136         Curl->protocol = malloc(pos1+1);
 137         if( Curl->protocol==NULL ) {
 138                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 139                 goto err_out;
 140         }
 141         strncpy(Curl->protocol, escfilename, pos1);
 142         Curl->protocol[pos1] = '\0';
 143
 144         // jump the "://"
 145         ptr1 += jumpSize;
 146         pos1 += jumpSize;
 147
 148         // check if a username:password is given
 149         ptr2 = strstr(ptr1, "@");
 150         ptr3 = strstr(ptr1, "/");
 151         if( ptr3!=NULL && ptr3<ptr2 ) {
 152                 // it isn't really a username but rather a part of the path
 153                 ptr2 = NULL;
 154         }
 155         if( ptr2!=NULL ) {
 156                 // We got something, at least a username...
 157                 int len = ptr2-ptr1;
 158                 Curl->username = malloc(len+1);
 159                 if( Curl->username==NULL ) {
 160                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 161                         goto err_out;
 162                 }
 163                 strncpy(Curl->username, ptr1, len);
 164                 Curl->username[len] = '\0';
 165
 166                 ptr3 = strstr(ptr1, ":");
 167                 if( ptr3!=NULL && ptr3<ptr2 ) {
 168                         // We also have a password
 169                         int len2 = ptr2-ptr3-1;
 170                         Curl->username[ptr3-ptr1]='\0';
 171                         Curl->password = malloc(len2+1);
 172                         if( Curl->password==NULL ) {
 173                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 174                                 goto err_out;
 175                         }
 176                         strncpy( Curl->password, ptr3+1, len2);
 177                         Curl->password[len2]='\0';
 178                         url_unescape_string(Curl->password, Curl->password);
 179                 }
 180                 url_unescape_string(Curl->username, Curl->username);
 181                 ptr1 = ptr2+1;
 182                 pos1 = ptr1-escfilename;
 183         }
 184
 185         // before looking for a port number check if we have an IPv6 type numeric address
 186         // in IPv6 URL the numeric address should be inside square braces.
 187         ptr2 = strstr(ptr1, "[");
 188         ptr3 = strstr(ptr1, "]");
 189         ptr4 = strstr(ptr1, "/");
 190         if( ptr2!=NULL && ptr3!=NULL && ptr2 < ptr3 && (!ptr4 || ptr4 > ptr3)) {
 191                 // we have an IPv6 numeric address
 192                 ptr1++;
 193                 pos1++;
 194                 ptr2 = ptr3;
 195                 v6addr = 1;
 196         } else {
 197                 ptr2 = ptr1;
 198
 199         }
 200
 201         // look if the port is given
 202         ptr2 = strstr(ptr2, ":");
 203         // If the : is after the first / it isn't the port
 204         ptr3 = strstr(ptr1, "/");
 205         if(ptr3 && ptr3 - ptr2 < 0) ptr2 = NULL;
 206         if( ptr2==NULL ) {
 207                 // No port is given
 208                 // Look if a path is given
 209                 if( ptr3==NULL ) {
 210                         // No path/filename
 211                         // So we have an URL like http://www.hostname.com
 212                         pos2 = strlen(escfilename);
 213                 } else {
 214                         // We have an URL like http://www.hostname.com/file.txt
 215                         pos2 = ptr3-escfilename;
 216                 }
 217         } else {
 218                 // We have an URL beginning like http://www.hostname.com:1212
 219                 // Get the port number
 220                 Curl->port = atoi(ptr2+1);
 221                 pos2 = ptr2-escfilename;
 222         }
 223         if( v6addr ) pos2--;
 224         // copy the hostname in the URL container
 225         Curl->hostname = malloc(pos2-pos1+1);
 226         if( Curl->hostname==NULL ) {
 227                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 228                 goto err_out;
 229         }
 230         strncpy(Curl->hostname, ptr1, pos2-pos1);
 231         Curl->hostname[pos2-pos1] = '\0';
 232
 233         // Look if a path is given
 234         ptr2 = strstr(ptr1, "/");
 235         if( ptr2!=NULL ) {
 236                 // A path/filename is given
 237                 // check if it's not a trailing '/'
 238                 if( strlen(ptr2)>1 ) {
 239                         // copy the path/filename in the URL container
 240                         Curl->file = strdup(ptr2);
 241                         if( Curl->file==NULL ) {
 242                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 243                                 goto err_out;
 244                         }
 245                 }
 246         }
 247         // Check if a filename was given or set, else set it with '/'
 248         if( Curl->file==NULL ) {
 249                 Curl->file = malloc(2);
 250                 if( Curl->file==NULL ) {
 251                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 252                         goto err_out;
 253                 }
 254                 strcpy(Curl->file, "/");
 255         }
 256
 257         noauth_len = make_noauth_url(Curl, NULL, 0);
 258         if (noauth_len > 0) {
 259                 noauth_len++;
 260                 Curl->noauth_url = malloc(noauth_len);
 261                 if (!Curl->noauth_url) {
 262                         mp_msg(MSGT_NETWORK, MSGL_FATAL, "Memory allocation failed.\n");
 263                         goto err_out;
 264                 }
 265                 make_noauth_url(Curl, Curl->noauth_url, noauth_len);
 266         }
 267
 268         free(escfilename);
 269         return Curl;
 270 err_out:
 271         free(escfilename);
 272         if (Curl) url_free(Curl);
 273         return NULL;
 274 }
 275
 276 void
 277 url_free(URL_t* url) {
 278         if(!url) return;
 279         free(url->url);
 280         free(url->protocol);
 281         free(url->hostname);
 282         free(url->file);
 283         free(url->username);
 284         free(url->password);
 285         free(url);
 286 }
 287
 288
 289 /* Replace escape sequences in an URL (or a part of an URL) */
 290 /* works like strcpy(), but without return argument,
 291    except that outbuf == inbuf is allowed */
 292 void
 293 url_unescape_string(char *outbuf, const char *inbuf)
 294 {
 295         unsigned char c,c1,c2;
 296         int i,len=strlen(inbuf);
 297         for (i=0;i<len;i++){
 298                 c = inbuf[i];
 299                 if (c == '%' && i<len-2) { //must have 2 more chars
 300                         c1 = toupper(inbuf[i+1]); // we need uppercase characters
 301                         c2 = toupper(inbuf[i+2]);
 302                         if (    ((c1>='0' && c1<='9') || (c1>='A' && c1<='F')) &&
 303                                 ((c2>='0' && c2<='9') || (c2>='A' && c2<='F')) ) {
 304                                 if (c1>='0' && c1<='9') c1-='0';
 305                                 else c1-='A'-10;
 306                                 if (c2>='0' && c2<='9') c2-='0';
 307                                 else c2-='A'-10;
 308                                 c = (c1<<4) + c2;
 309                                 i=i+2; //only skip next 2 chars if valid esc
 310                         }
 311                 }
 312                 *outbuf++ = c;
 313         }
 314         *outbuf++='\0'; //add nullterm to string
 315 }
 316
 317 static void
 318 url_escape_string_part(char *outbuf, const char *inbuf) {
 319         unsigned char c,c1,c2;
 320         int i,len=strlen(inbuf);
 321
 322         for  (i=0;i<len;i++) {
 323                 c = inbuf[i];
 324                 if ((c=='%') && i<len-2 ) { //need 2 more characters
 325                     c1=toupper(inbuf[i+1]); c2=toupper(inbuf[i+2]); // need uppercase chars
 326                    } else {
 327                     c1=129; c2=129; //not escape chars
 328                    }
 329
 330                 if(     (c >= 'A' && c <= 'Z') ||
 331                         (c >= 'a' && c <= 'z') ||
 332                         (c >= '0' && c <= '9') ||
 333                         (c >= 0x7f)) {
 334                         *outbuf++ = c;
 335                 } else if ( c=='%' && ((c1 >= '0' && c1 <= '9') || (c1 >= 'A' && c1 <= 'F')) &&
 336                            ((c2 >= '0' && c2 <= '9') || (c2 >= 'A' && c2 <= 'F'))) {
 337                                                               // check if part of an escape sequence
 338                             *outbuf++=c;                      // already
 339
 340                                                               // dont escape again
 341                             mp_tmsg(MSGT_NETWORK,MSGL_ERR,"String appears to be already escaped in url_escape %c%c1%c2\n",c,c1,c2);
 342                                                               // error as this should not happen against RFC 2396
 343                                                               // to escape a string twice
 344                 } else {
 345                         /* all others will be escaped */
 346                         c1 = ((c & 0xf0) >> 4);
 347                         c2 = (c & 0x0f);
 348                         if (c1 < 10) c1+='0';
 349                         else c1+='A'-10;
 350                         if (c2 < 10) c2+='0';
 351                         else c2+='A'-10;
 352                         *outbuf++ = '%';
 353                         *outbuf++ = c1;
 354                         *outbuf++ = c2;
 355                 }
 356         }
 357         *outbuf++='\0';
 358 }
 359
 360 /* Replace specific characters in the URL string by an escape sequence */
 361 /* works like strcpy(), but without return argument */
 362 void
 363 url_escape_string(char *outbuf, const char *inbuf) {
 364         unsigned char c;
 365         int i = 0,j,len = strlen(inbuf);
 366         char* tmp,*unesc = NULL, *in;
 367
 368         // Look if we have an ip6 address, if so skip it there is
 369         // no need to escape anything in there.
 370         tmp = strstr(inbuf,"://[");
 371         if(tmp) {
 372                 tmp = strchr(tmp+4,']');
 373                 if(tmp && (tmp[1] == '/' || tmp[1] == ':' ||
 374                            tmp[1] == '\0')) {
 375                         i = tmp+1-inbuf;
 376                         strncpy(outbuf,inbuf,i);
 377                         outbuf += i;
 378                         tmp = NULL;
 379                 }
 380         }
 381
 382         tmp = NULL;
 383         while(i < len) {
 384                 // look for the next char that must be kept
 385                 for  (j=i;j<len;j++) {
 386                         c = inbuf[j];
 387                         if(c=='-' || c=='_' || c=='.' || c=='!' || c=='~' ||    /* mark characters */
 388                            c=='*' || c=='\'' || c=='(' || c==')' ||             /* do not touch escape character */
 389                            c==';' || c=='/' || c=='?' || c==':' || c=='@' ||    /* reserved characters */
 390                            c=='&' || c=='=' || c=='+' || c=='$' || c==',')      /* see RFC 2396 */
 391                                 break;
 392                 }
 393                 // we are on a reserved char, write it out
 394                 if(j == i) {
 395                         *outbuf++ = c;
 396                         i++;
 397                         continue;
 398                 }
 399                 // we found one, take that part of the string
 400                 if(j < len) {
 401                         if(!tmp) tmp = malloc(len+1);
 402                         strncpy(tmp,inbuf+i,j-i);
 403                         tmp[j-i] = '\0';
 404                         in = tmp;
 405                 } else // take the rest of the string
 406                         in = (char*)inbuf+i;
 407
 408                 if(!unesc) unesc = malloc(len+1);
 409                 // unescape first to avoid escaping escape
 410                 url_unescape_string(unesc,in);
 411                 // then escape, including mark and other reserved chars
 412                 // that can come from escape sequences
 413                 url_escape_string_part(outbuf,unesc);
 414                 outbuf += strlen(outbuf);
 415                 i += strlen(in);
 416         }
 417         *outbuf = '\0';
 418         free(tmp);
 419         free(unesc);
 420 }
 421
 422 #ifdef URL_DEBUG
 423 void
 424 url_debug(const URL_t *url) {
 425         if( url==NULL ) {
 426                 printf("URL pointer NULL\n");
 427                 return;
 428         }
 429         if( url->url!=NULL ) {
 430                 printf("url=%s\n", url->url );
 431         }
 432         if( url->protocol!=NULL ) {
 433                 printf("protocol=%s\n", url->protocol );
 434         }
 435         if( url->hostname!=NULL ) {
 436                 printf("hostname=%s\n", url->hostname );
 437         }
 438         printf("port=%d\n", url->port );
 439         if( url->file!=NULL ) {
 440                 printf("file=%s\n", url->file );
 441         }
 442         if( url->username!=NULL ) {
 443                 printf("username=%s\n", url->username );
 444         }
 445         if( url->password!=NULL ) {
 446                 printf("password=%s\n", url->password );
 447         }
 448 }
 449 #endif /* URL_DEBUG */