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 URL_t*
  61 url_new(const char* url) {
  62         int pos1, pos2,v6addr = 0;
  63         URL_t* Curl = NULL;
  64         char *escfilename=NULL;
  65         char *ptr1=NULL, *ptr2=NULL, *ptr3=NULL, *ptr4=NULL;
  66         int jumpSize = 3;
  67
  68         if( url==NULL ) return NULL;
  69
  70         if (strlen(url) > (SIZE_MAX / 3 - 1)) {
  71                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
  72                 goto err_out;
  73         }
  74         escfilename=malloc(strlen(url)*3+1);
  75         if (!escfilename ) {
  76                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
  77                 goto err_out;
  78         }
  79
  80         // Create the URL container
  81         Curl = malloc(sizeof(URL_t));
  82         if( Curl==NULL ) {
  83                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
  84                 goto err_out;
  85         }
  86
  87         // Initialisation of the URL container members
  88         memset( Curl, 0, sizeof(URL_t) );
  89
  90         url_escape_string(escfilename,url);
  91
  92         // Copy the url in the URL container
  93         Curl->url = strdup(escfilename);
  94         if( Curl->url==NULL ) {
  95                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
  96                 goto err_out;
  97         }
  98         mp_msg(MSGT_OPEN,MSGL_V,"Filename for url is now %s\n",escfilename);
  99
 100         // extract the protocol
 101         ptr1 = strstr(escfilename, "://");
 102         if( ptr1==NULL ) {
 103                 // Check for a special case: "sip:" (without "//"):
 104                 if (strstr(escfilename, "sip:") == escfilename) {
 105                         ptr1 = (char *)&url[3]; // points to ':'
 106                         jumpSize = 1;
 107                 } else {
 108                         mp_msg(MSGT_NETWORK,MSGL_V,"Not an URL!\n");
 109                         goto err_out;
 110                 }
 111         }
 112         pos1 = ptr1-escfilename;
 113         Curl->protocol = malloc(pos1+1);
 114         if( Curl->protocol==NULL ) {
 115                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 116                 goto err_out;
 117         }
 118         strncpy(Curl->protocol, escfilename, pos1);
 119         Curl->protocol[pos1] = '\0';
 120
 121         // jump the "://"
 122         ptr1 += jumpSize;
 123         pos1 += jumpSize;
 124
 125         // check if a username:password is given
 126         ptr2 = strstr(ptr1, "@");
 127         ptr3 = strstr(ptr1, "/");
 128         if( ptr3!=NULL && ptr3<ptr2 ) {
 129                 // it isn't really a username but rather a part of the path
 130                 ptr2 = NULL;
 131         }
 132         if( ptr2!=NULL ) {
 133                 // We got something, at least a username...
 134                 int len = ptr2-ptr1;
 135                 Curl->username = malloc(len+1);
 136                 if( Curl->username==NULL ) {
 137                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 138                         goto err_out;
 139                 }
 140                 strncpy(Curl->username, ptr1, len);
 141                 Curl->username[len] = '\0';
 142
 143                 ptr3 = strstr(ptr1, ":");
 144                 if( ptr3!=NULL && ptr3<ptr2 ) {
 145                         // We also have a password
 146                         int len2 = ptr2-ptr3-1;
 147                         Curl->username[ptr3-ptr1]='\0';
 148                         Curl->password = malloc(len2+1);
 149                         if( Curl->password==NULL ) {
 150                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 151                                 goto err_out;
 152                         }
 153                         strncpy( Curl->password, ptr3+1, len2);
 154                         Curl->password[len2]='\0';
 155                 }
 156                 ptr1 = ptr2+1;
 157                 pos1 = ptr1-escfilename;
 158         }
 159
 160         // before looking for a port number check if we have an IPv6 type numeric address
 161         // in IPv6 URL the numeric address should be inside square braces.
 162         ptr2 = strstr(ptr1, "[");
 163         ptr3 = strstr(ptr1, "]");
 164         ptr4 = strstr(ptr1, "/");
 165         if( ptr2!=NULL && ptr3!=NULL && ptr2 < ptr3 && (!ptr4 || ptr4 > ptr3)) {
 166                 // we have an IPv6 numeric address
 167                 ptr1++;
 168                 pos1++;
 169                 ptr2 = ptr3;
 170                 v6addr = 1;
 171         } else {
 172                 ptr2 = ptr1;
 173
 174         }
 175
 176         // look if the port is given
 177         ptr2 = strstr(ptr2, ":");
 178         // If the : is after the first / it isn't the port
 179         ptr3 = strstr(ptr1, "/");
 180         if(ptr3 && ptr3 - ptr2 < 0) ptr2 = NULL;
 181         if( ptr2==NULL ) {
 182                 // No port is given
 183                 // Look if a path is given
 184                 if( ptr3==NULL ) {
 185                         // No path/filename
 186                         // So we have an URL like http://www.hostname.com
 187                         pos2 = strlen(escfilename);
 188                 } else {
 189                         // We have an URL like http://www.hostname.com/file.txt
 190                         pos2 = ptr3-escfilename;
 191                 }
 192         } else {
 193                 // We have an URL beginning like http://www.hostname.com:1212
 194                 // Get the port number
 195                 Curl->port = atoi(ptr2+1);
 196                 pos2 = ptr2-escfilename;
 197         }
 198         if( v6addr ) pos2--;
 199         // copy the hostname in the URL container
 200         Curl->hostname = malloc(pos2-pos1+1);
 201         if( Curl->hostname==NULL ) {
 202                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 203                 goto err_out;
 204         }
 205         strncpy(Curl->hostname, ptr1, pos2-pos1);
 206         Curl->hostname[pos2-pos1] = '\0';
 207
 208         // Look if a path is given
 209         ptr2 = strstr(ptr1, "/");
 210         if( ptr2!=NULL ) {
 211                 // A path/filename is given
 212                 // check if it's not a trailing '/'
 213                 if( strlen(ptr2)>1 ) {
 214                         // copy the path/filename in the URL container
 215                         Curl->file = strdup(ptr2);
 216                         if( Curl->file==NULL ) {
 217                                 mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 218                                 goto err_out;
 219                         }
 220                 }
 221         }
 222         // Check if a filename was given or set, else set it with '/'
 223         if( Curl->file==NULL ) {
 224                 Curl->file = malloc(2);
 225                 if( Curl->file==NULL ) {
 226                         mp_tmsg(MSGT_NETWORK,MSGL_FATAL,"Memory allocation failed.\n");
 227                         goto err_out;
 228                 }
 229                 strcpy(Curl->file, "/");
 230         }
 231
 232         free(escfilename);
 233         return Curl;
 234 err_out:
 235         if (escfilename) free(escfilename);
 236         if (Curl) url_free(Curl);
 237         return NULL;
 238 }
 239
 240 void
 241 url_free(URL_t* url) {
 242         if(!url) return;
 243         if(url->url) free(url->url);
 244         if(url->protocol) free(url->protocol);
 245         if(url->hostname) free(url->hostname);
 246         if(url->file) free(url->file);
 247         if(url->username) free(url->username);
 248         if(url->password) free(url->password);
 249         free(url);
 250 }
 251
 252
 253 /* Replace escape sequences in an URL (or a part of an URL) */
 254 /* works like strcpy(), but without return argument */
 255 void
 256 url_unescape_string(char *outbuf, const char *inbuf)
 257 {
 258         unsigned char c,c1,c2;
 259         int i,len=strlen(inbuf);
 260         for (i=0;i<len;i++){
 261                 c = inbuf[i];
 262                 if (c == '%' && i<len-2) { //must have 2 more chars
 263                         c1 = toupper(inbuf[i+1]); // we need uppercase characters
 264                         c2 = toupper(inbuf[i+2]);
 265                         if (    ((c1>='0' && c1<='9') || (c1>='A' && c1<='F')) &&
 266                                 ((c2>='0' && c2<='9') || (c2>='A' && c2<='F')) ) {
 267                                 if (c1>='0' && c1<='9') c1-='0';
 268                                 else c1-='A'-10;
 269                                 if (c2>='0' && c2<='9') c2-='0';
 270                                 else c2-='A'-10;
 271                                 c = (c1<<4) + c2;
 272                                 i=i+2; //only skip next 2 chars if valid esc
 273                         }
 274                 }
 275                 *outbuf++ = c;
 276         }
 277         *outbuf++='\0'; //add nullterm to string
 278 }
 279
 280 static void
 281 url_escape_string_part(char *outbuf, const char *inbuf) {
 282         unsigned char c,c1,c2;
 283         int i,len=strlen(inbuf);
 284
 285         for  (i=0;i<len;i++) {
 286                 c = inbuf[i];
 287                 if ((c=='%') && i<len-2 ) { //need 2 more characters
 288                     c1=toupper(inbuf[i+1]); c2=toupper(inbuf[i+2]); // need uppercase chars
 289                    } else {
 290                     c1=129; c2=129; //not escape chars
 291                    }
 292
 293                 if(     (c >= 'A' && c <= 'Z') ||
 294                         (c >= 'a' && c <= 'z') ||
 295                         (c >= '0' && c <= '9') ||
 296                         (c >= 0x7f)) {
 297                         *outbuf++ = c;
 298                 } else if ( c=='%' && ((c1 >= '0' && c1 <= '9') || (c1 >= 'A' && c1 <= 'F')) &&
 299                            ((c2 >= '0' && c2 <= '9') || (c2 >= 'A' && c2 <= 'F'))) {
 300                                                               // check if part of an escape sequence
 301                             *outbuf++=c;                      // already
 302
 303                                                               // dont escape again
 304                             mp_tmsg(MSGT_NETWORK,MSGL_ERR,"String appears to be already escaped in url_escape %c%c1%c2\n",c,c1,c2);
 305                                                               // error as this should not happen against RFC 2396
 306                                                               // to escape a string twice
 307                 } else {
 308                         /* all others will be escaped */
 309                         c1 = ((c & 0xf0) >> 4);
 310                         c2 = (c & 0x0f);
 311                         if (c1 < 10) c1+='0';
 312                         else c1+='A'-10;
 313                         if (c2 < 10) c2+='0';
 314                         else c2+='A'-10;
 315                         *outbuf++ = '%';
 316                         *outbuf++ = c1;
 317                         *outbuf++ = c2;
 318                 }
 319         }
 320         *outbuf++='\0';
 321 }
 322
 323 /* Replace specific characters in the URL string by an escape sequence */
 324 /* works like strcpy(), but without return argument */
 325 void
 326 url_escape_string(char *outbuf, const char *inbuf) {
 327         unsigned char c;
 328         int i = 0,j,len = strlen(inbuf);
 329         char* tmp,*unesc = NULL, *in;
 330
 331         // Look if we have an ip6 address, if so skip it there is
 332         // no need to escape anything in there.
 333         tmp = strstr(inbuf,"://[");
 334         if(tmp) {
 335                 tmp = strchr(tmp+4,']');
 336                 if(tmp && (tmp[1] == '/' || tmp[1] == ':' ||
 337                            tmp[1] == '\0')) {
 338                         i = tmp+1-inbuf;
 339                         strncpy(outbuf,inbuf,i);
 340                         outbuf += i;
 341                         tmp = NULL;
 342                 }
 343         }
 344
 345         tmp = NULL;
 346         while(i < len) {
 347                 // look for the next char that must be kept
 348                 for  (j=i;j<len;j++) {
 349                         c = inbuf[j];
 350                         if(c=='-' || c=='_' || c=='.' || c=='!' || c=='~' ||    /* mark characters */
 351                            c=='*' || c=='\'' || c=='(' || c==')' ||             /* do not touch escape character */
 352                            c==';' || c=='/' || c=='?' || c==':' || c=='@' ||    /* reserved characters */
 353                            c=='&' || c=='=' || c=='+' || c=='$' || c==',')      /* see RFC 2396 */
 354                                 break;
 355                 }
 356                 // we are on a reserved char, write it out
 357                 if(j == i) {
 358                         *outbuf++ = c;
 359                         i++;
 360                         continue;
 361                 }
 362                 // we found one, take that part of the string
 363                 if(j < len) {
 364                         if(!tmp) tmp = malloc(len+1);
 365                         strncpy(tmp,inbuf+i,j-i);
 366                         tmp[j-i] = '\0';
 367                         in = tmp;
 368                 } else // take the rest of the string
 369                         in = (char*)inbuf+i;
 370
 371                 if(!unesc) unesc = malloc(len+1);
 372                 // unescape first to avoid escaping escape
 373                 url_unescape_string(unesc,in);
 374                 // then escape, including mark and other reserved chars
 375                 // that can come from escape sequences
 376                 url_escape_string_part(outbuf,unesc);
 377                 outbuf += strlen(outbuf);
 378                 i += strlen(in);
 379         }
 380         *outbuf = '\0';
 381         if(tmp) free(tmp);
 382         if(unesc) free(unesc);
 383 }
 384
 385 #ifdef URL_DEBUG
 386 void
 387 url_debug(const URL_t *url) {
 388         if( url==NULL ) {
 389                 printf("URL pointer NULL\n");
 390                 return;
 391         }
 392         if( url->url!=NULL ) {
 393                 printf("url=%s\n", url->url );
 394         }
 395         if( url->protocol!=NULL ) {
 396                 printf("protocol=%s\n", url->protocol );
 397         }
 398         if( url->hostname!=NULL ) {
 399                 printf("hostname=%s\n", url->hostname );
 400         }
 401         printf("port=%d\n", url->port );
 402         if( url->file!=NULL ) {
 403                 printf("file=%s\n", url->file );
 404         }
 405         if( url->username!=NULL ) {
 406                 printf("username=%s\n", url->username );
 407         }
 408         if( url->password!=NULL ) {
 409                 printf("password=%s\n", url->password );
 410         }
 411 }
 412 #endif /* URL_DEBUG */