#include <gnunet_util_lib.h>
#include <microhttpd.h>
#include <curl/curl.h>
+#include <regex.h>
#include "gns_proxy_proto.h"
#include "gns.h"
#define BUF_WAIT_FOR_CURL 0
#define BUF_WAIT_FOR_MHD 1
+#define HTML_HDR_CONTENT "Content-Type: text/html\r\n"
+#define RE_DOTPLUS "<a href=\"http://(([A-Za-z]+[.])+)([+])"
+#define RE_N_MATCHES 4
+
+#define HTTP_PORT 80
+#define HTTPS_PORT 443
struct ProxyCurlTask
{
int download_successful;
int download_error;
struct MHD_Connection *connection;
+ int parse_content;
+ int is_postprocessing;
+
+ GNUNET_SCHEDULER_TaskIdentifier pp_task;
+
+ char pp_buf[256];
};
struct ProxyCurlTask *ctasks_head;
struct ProxyCurlTask *ctasks_tail;
+static regex_t re_dotplus;
+
+/**
+ * Read HTTP request header field 'Host'
+ *
+ * @param cls buffer to write to
+ * @param kind value kind
+ * @param key field key
+ * @param value field value
+ * @return MHD_NO when Host found
+ */
static int
con_val_iter (void *cls,
enum MHD_ValueKind kind,
}
+/**
+ * Check HTTP response header for mime
+ *
+ * @param buffer curl buffer
+ * @param size curl blocksize
+ * @param nmemb curl blocknumber
+ * @param cls handle
+ * @return size of read bytes
+ */
+static size_t
+curl_check_hdr (void *buffer, size_t size, size_t nmemb, void *cls)
+{
+ size_t bytes = size * nmemb;
+ struct ProxyCurlTask *ctask = cls;
+ char hdr[bytes+1];
+
+ memcpy (hdr, buffer, bytes);
+ hdr[bytes] = '\0';
+
+ if (0 == strcmp (hdr, HTML_HDR_CONTENT))
+ {
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Got HTML HTTP response header\n");
+ ctask->parse_content = GNUNET_YES;
+ }
+
+ return bytes;
+}
+
/**
* schedule mhd
*/
}
+static void
+postprocess_name (void *cls, const struct GNUNET_SCHEDULER_TaskContext *tc)
+{
+ struct ProxyCurlTask *ctask = cls;
+ char tmp[strlen(ctask->pp_buf)];
+
+ sprintf ( tmp, "<a href=http://%sxxx.gnunet", ctask->pp_buf);
+
+ strcpy (ctask->pp_buf, tmp);
+
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Created %s\n", tmp);
+
+ run_httpd();
+ //MHD_run (httpd);
+}
+
/**
* Callback for MHD response
*
{
struct ProxyCurlTask *ctask = cls;
ssize_t copied = 0;
+ size_t bytes_to_copy;
+ int nomatch;
+ char *hostptr;
+ regmatch_t m[RE_N_MATCHES];
+
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "MHD: content cb\n");
if (ctask->download_successful &&
(ctask->buf_status == BUF_WAIT_FOR_CURL))
}
if ( ctask->buf_status == BUF_WAIT_FOR_CURL )
- {
return 0;
+
+ bytes_to_copy = ctask->bytes_in_buffer;
+
+ if (ctask->parse_content == GNUNET_YES)
+ {
+
+ GNUNET_log ( GNUNET_ERROR_TYPE_DEBUG,
+ "MHD: We need to parse the HTML %s\n", ctask->buffer_ptr);
+
+ nomatch = regexec ( &re_dotplus, ctask->buffer_ptr, RE_N_MATCHES, m, 0);
+
+ if (!nomatch)
+ {
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "MHD RE: Match\n");
+
+ GNUNET_assert (m[1].rm_so != -1);
+
+ hostptr = ctask->buffer_ptr+m[1].rm_so;
+
+ if (m[0].rm_so > 0)
+ {
+ bytes_to_copy = m[0].rm_so;
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Copying %d bytes.\n", m[0].rm_so);
+
+
+ }
+ else
+ {
+ if (ctask->is_postprocessing == GNUNET_YES)
+ {
+
+ /*Done?*/
+ if ( 0 == strcmp (ctask->pp_buf, "") )
+ return 0;
+
+ ctask->is_postprocessing = GNUNET_NO;
+
+ ctask->bytes_in_buffer -= m[0].rm_eo;//(m[1].rm_eo-m[1].rm_so);
+ ctask->buffer_ptr += m[0].rm_eo;//(m[1].rm_eo-m[1].rm_so);
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Skipping next %d bytes in buffer\n", m[0].rm_eo);
+
+ if ( strlen (ctask->pp_buf) <= max )
+ {
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Copying postprocessed %s.\n", ctask->pp_buf);
+ memcpy ( buf, ctask->pp_buf, strlen (ctask->pp_buf) );
+ GNUNET_log (GNUNET_ERROR_TYPE_DEBUG,
+ "Done %s.\n", buf);
+ ctask->is_postprocessing = GNUNET_NO;
+ return strlen (ctask->pp_buf);
+ }
+
+ return 0;
+ }
+
+ memset (ctask->pp_buf, 0, sizeof(ctask->pp_buf));
+ memcpy (ctask->pp_buf, hostptr, (m[1].rm_eo-m[1].rm_so));
+
+ ctask->is_postprocessing = GNUNET_YES;
+
+ postprocess_name(ctask, NULL);
+ //ctask->pp_task = GNUNET_SCHEDULER_add_now (&postprocess_name, ctask);
+
+ return 0;
+ }
+ }
}
- if ( ctask->bytes_in_buffer > max )
+ if ( bytes_to_copy > max )
{
GNUNET_log ( GNUNET_ERROR_TYPE_DEBUG,
"MHD: buffer in response too small! (%s)\n",
{
GNUNET_log ( GNUNET_ERROR_TYPE_DEBUG,
"MHD: copying %d bytes to mhd response at offset %d\n",
- ctask->bytes_in_buffer, pos);
+ bytes_to_copy, pos);
- memcpy ( buf, ctask->buffer_ptr, ctask->bytes_in_buffer );
- copied = ctask->bytes_in_buffer;
- ctask->buf_status = BUF_WAIT_FOR_CURL;
- ctask->buffer_ptr = ctask->buffer;
- curl_easy_pause (ctask->curl, CURLPAUSE_CONT);
+ memcpy ( buf, ctask->buffer_ptr, bytes_to_copy );
+ copied = bytes_to_copy;
+ if (bytes_to_copy < ctask->bytes_in_buffer)
+ {
+ ctask->bytes_in_buffer -= bytes_to_copy;
+ ctask->buffer_ptr += bytes_to_copy;
+ }
+ else
+ {
+ ctask->bytes_in_buffer = 0;
+ ctask->buf_status = BUF_WAIT_FOR_CURL;
+ ctask->buffer_ptr = ctask->buffer;
+ curl_easy_pause (ctask->curl, CURLPAUSE_CONT);
+ }
}
return copied;
ctask->connection = con;
ctask->buf_status = BUF_WAIT_FOR_CURL;
ctask->bytes_in_buffer = 0;
+ ctask->parse_content = GNUNET_NO;
+ curl_easy_setopt (ctask->curl, CURLOPT_HEADERFUNCTION, &curl_check_hdr);
+ curl_easy_setopt (ctask->curl, CURLOPT_HEADERDATA, ctask);
curl_easy_setopt (ctask->curl, CURLOPT_WRITEFUNCTION, &callback_download);
curl_easy_setopt (ctask->curl, CURLOPT_WRITEDATA, ctask);
curl_easy_setopt (ctask->curl, CURLOPT_FOLLOWLOCATION, 1);
}
}
+/**
+ * Compiles a regex for us
+ *
+ * @param re ptr to re struct
+ * @param rt the expression to compile
+ * @return 0 on success
+ */
+static int
+compile_regex (regex_t *re, const char* rt)
+{
+ int status;
+ char err[1024];
+
+ status = regcomp (re, rt, REG_EXTENDED|REG_NEWLINE);
+ if (status)
+ {
+ regerror (status, re, err, 1024);
+ GNUNET_log (GNUNET_ERROR_TYPE_ERROR,
+ "Regex error compiling '%s': %s\n", rt, err);
+ return 1;
+ }
+ return 0;
+}
+
/**
* Main function that will be run
*
{
struct sockaddr_in sa;
+ compile_regex (&re_dotplus, (char*) RE_DOTPLUS);
+
memset (&sa, 0, sizeof (sa));
sa.sin_family = AF_INET;
sa.sin_port = htons (port);