David
in order to send the application the info below, you have to navigate
the ndpi flow pointer and read it. In general however this mechanism
does not work with HTTP 1.1 as nDPI identifies only the first
URL/request of the TCP session, as this is enough for DPI.

Regards Luca


On 07/29/2014 02:08 AM, David Britt wrote:
> I am trying to capture the HTTP Content type and Referer for a
> specific HTTP flow - browsing through the nDPI source I notice that
> there seem to be a bunch of structure defined in ndpi_struct.h as
> follows:
>
>   struct ndpi_int_one_line_struct line[NDPI_MAX_PARSE_LINES_PER_PACKET];
>   struct ndpi_int_one_line_struct
> unix_line[NDPI_MAX_PARSE_LINES_PER_PACKET];
>   struct ndpi_int_one_line_struct host_line;
>   struct ndpi_int_one_line_struct forwarded_line;
>   struct ndpi_int_one_line_struct referer_line;
>   struct ndpi_int_one_line_struct content_line;
>   struct ndpi_int_one_line_struct accept_line;
>   struct ndpi_int_one_line_struct user_agent_line;
>   struct ndpi_int_one_line_struct http_url_name;
>   struct ndpi_int_one_line_struct http_encoding;
>   struct ndpi_int_one_line_struct http_transfer_encoding;
>   struct ndpi_int_one_line_struct http_contentlen;
>   struct ndpi_int_one_line_struct http_cookie;
>   struct ndpi_int_one_line_struct http_x_session_type;
>   struct ndpi_int_one_line_struct server_line;
>   struct ndpi_int_one_line_struct http_method;
>   struct ndpi_int_one_line_struct http_response;
>
> It seems that the host_line is copied to the host_server_name in
> http.c to give the URL, is it possible to also grab other information
> such as the content type (is this content_line as above?) and referer
> (referer_line as above?).  
>
> It is not obvious to me exactly how the URL is actually captured and
> passed so hopefully there is simple way to achieve this without
> wholesale modification.
>
> Before I go ahead and try and modify my code to include this I thought
> it might be worth asking to make sure I am not missing something obvious.
>
> Ultimately I am trying to capture the referer and content type to
> potentially allow only display of originating URLs. That is, if I open
> a web page that contains many embedded URLs (for example
> advertisements) I only want to record the URL the user actually
> visited and ignore any embedded URLs in the page that currently are
> shown as separate flows with their own URL. 
>
> Of course, if I can work this out in the interim I will be sure to
> post - who knows there just may be someone else with a similar
> requirement.
>
> Thanks,
>
> Dave.
>
>
> _______________________________________________
> Ntop-misc mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop-misc

_______________________________________________
Ntop-misc mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-misc

Reply via email to