Commit 59991b0f authored by robertdavidgraham's avatar robertdavidgraham
Browse files

#86 --capture html

parent 148b58bb

doc/masscan.8.markdown

+7 −0
Original line number Diff line number Diff line
@@ -255,6 +255,13 @@ one port. 
    contents of the BASE64 encoded string are decoded, then used as the hello

    string that greets the server.



  * `--capture <type>` or `--nocapture <type>`: when doing banners (`--banner`), this

    determines what to capture from the banners. By default, only the TITLE field from

	HTML documents is captured, to get the entire document, use `--capture html`.

	By default, the entire certificate from SSL is captured, to disable this, use

	`--nocapture cert`. Currently, only the values `html` and `cert` are currently

	supported for this option, but many more will be added in the future.

    



## CONFIGURATION FILE FORMAT

src/proto-http.c

+20 −5
Original line number Diff line number Diff line
@@ -346,19 +346,31 @@ http_parse( 
        }

        break;

    case CONTENT:

        {

            unsigned next = i;



            id = smack_search_next(

                                   banner1->html_fields,

                                   &state2,

                                   px, &i, (unsigned)length);

            i--;

                                   px, &next, (unsigned)length);



            if (banner1->is_capture_html) {

                banout_append(banout, PROTO_HTML_FULL, &px[i], next-i);

            }



            if (id != SMACK_NOT_FOUND) {

                //field_name(banout, id, html_fields);

                //banout_append_char(banout, PROTO_HTML_TITLE, ':');

                state = CONTENT_TAG;

            }



            i = next - 1;

        }

        break;

    case CONTENT_TAG:

        for (; i<length; i++) {

            if (banner1->is_capture_html) {

                banout_append_char(banout, PROTO_HTML_FULL, px[i]);

            }



            if (px[i] == '>') {

                state = CONTENT_FIELD;

                break;
@@ -366,6 +378,9 @@ http_parse( 
        }

        break;

    case CONTENT_FIELD:

        if (banner1->is_capture_html) {

            banout_append_char(banout, PROTO_HTML_FULL, px[i]);

        }

        if (px[i] == '<')

            state = CONTENT;

        else {