Next: , Previous: , Up: Steps in Implementing an Insecure HTTP Client   [Index]

1.3.1.2 Implementing the HTTP Client Function ‘parse_url’

The main routine is invoked with a URL of the form 

http://www.server.com/path/to/document.html

You need to separate the host and the path using a utility routine parse_url, shown in Listing 1-2. 

/**
 * Accept a well-formed URL (e.g. http://www.company.com/index.html) and return
 * pointers to the host part and the path part. Note that this function
 * modifies the uri itself as well. It returns 0 on success, -1 if the URL is
 * found to be malformed in any way.
 */

int parse_url( char *uri, char **host, char **path )
{
  char *pos;

  pos = strstr( uri, "//" );

  if ( !pos )
    {
      return -1;
    }

  *host = pos + 2;

  pos = strchr( *host, '/' );

  if ( !pos )
    {
      *path = NULL;
    }
  else
    {
      *pos = '\0';
      *path = pos + 1;
    }

  return 0;
}

Listing 1.2: "http.c" parse_url

You scan through the URL, looking for the delimiters ‘//’ and ‘/’ and replace them with null-terminators so that the caller can treat them as C strings. The calling function passes in two pointers to pointers; these should be null when the function starts and will be modified to point into the ‘uri’ string, which came from argv.