------------------------------------------------------------------------
r17854 | luis | 2010-06-06 15:47:34 +0200 (Sun, 06 Jun 2010) | 1 line
Removed a few things from the documentation that were rerefenced but not yet implemented
------------------------------------------------------------------------
r17855 | luis | 2010-06-06 15:53:45 +0200 (Sun, 06 Jun 2010) | 1 line
Fixed a bug that caused ARP Requests to be sent no matter what the user supplied as the ARP operation code
------------------------------------------------------------------------
r17856 | luis | 2010-06-06 16:23:47 +0200 (Sun, 06 Jun 2010) | 1 line
Fixed a bug that produced a buffer overflow when non-RFC-compliant ICMP types were set. Now, if user supplies a weird ICMP type, a regular 8-byte ICMP header will be sent
------------------------------------------------------------------------
r17858 | luis | 2010-06-06 22:08:13 +0200 (Sun, 06 Jun 2010) | 1 line
Added a function (and a helper) so the number parsing in Nping is consistent. The new function, parse_unsigned_number() may parse 8, 16, or 32 bit unsigned integers, specified in hexadecimal, octal or decimal. Alternatively, source string may be the special value <<rand>> or <<random>>, and the function will generate a random number of the desired length. Function validate_number_spec() is needed due to the limitations of strtoul() that doesn't report errors on things like 1zzzppp, and gives undesired results when a minus sign is supplied. validate_number_spec() performs various checks, trying to detect bogus input. However, if something is not catched by it, strtoul() may take care of it. I've been testing both functions for hours, even with automatically generated input and I haven't been able to find any bug, so I think its safe to start using this.
------------------------------------------------------------------------
r17859 | luis | 2010-06-06 23:43:04 +0200 (Sun, 06 Jun 2010) | 1 line
Added the possibility to supply special value <<rand>> in MAC address specification
------------------------------------------------------------------------
r17860 | luis | 2010-06-07 00:01:30 +0200 (Mon, 07 Jun 2010) | 1 line
Added the possibility to specify the special value <<broadcast>> so the address FF:FF:FF:FF:FF:FF is set. Also, fixed move string length check so the function does not return OP_FAILURE before checking for these special values
------------------------------------------------------------------------
r17866 | luis | 2010-06-07 01:05:31 +0200 (Mon, 07 Jun 2010) | 1 line
Implemented function getRandomTextPayload() to support <<random>> value in payload string specification
------------------------------------------------------------------------
r17867 | luis | 2010-06-07 01:08:29 +0200 (Mon, 07 Jun 2010) | 1 line
Lots of changes in Nping's command-line argument parser to use the new parse_unsigned_number() function. Now virtually all parameters may be specified in hexadecimal, octal or decimal base, or even using the special value <<random>>
------------------------------------------------------------------------
r17868 | luis | 2010-06-07 01:12:02 +0200 (Mon, 07 Jun 2010) | 1 line
Removed useless var declarations. There is already an aux32 var declared at the start of the function
------------------------------------------------------------------------
r17869 | luis | 2010-06-07 01:24:00 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed a bug that caused the SYN flag to be set even when the user had specified --flags 0x00.
------------------------------------------------------------------------
r17870 | luis | 2010-06-07 01:27:33 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed another bug in flag parsing. When an empty string was being passed, the SYN bit was being set. Now if an empty string is passed, no flags will be set
------------------------------------------------------------------------
r17873 | luis | 2010-06-07 13:12:36 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed a bug that caused -g to be ignored in TCP-Connect or UDP unprivileged modes. Now, providing the user has enough privileges on his system, the specified source port will be set, at least in the first packet send to each target. I've tested this on Linux and the OS complains when Nping attempts to use the same source port twice for the same target IP address in TCP-Connect mode. For this reason, a new warning was added so the users know that it will probably fail.
------------------------------------------------------------------------
r17875 | luis | 2010-06-07 13:25:04 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed a bug that caused message <<Operation now in progress>> to be displayed in TCP-Connect mode when RST packets were being received. The problem was that errno was being used and by the time it was referenced, its value had changed so the wrong message was being printed. Now, the error code is obtained from nsock trough a call to nse_errorcode() and the correct message is displayed: <<Connection refused>>
------------------------------------------------------------------------
r17876 | luis | 2010-06-07 13:29:04 +0200 (Mon, 07 Jun 2010) | 1 line
Minor fix
------------------------------------------------------------------------
r17877 | luis | 2010-06-07 13:49:52 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed a bug that caused Nping to fail when the first supplied target was not resolvable (eg: nping bogushost.fkz scanme.insecure.com tcpdump.com).
------------------------------------------------------------------------
r17878 | luis | 2010-06-07 14:56:02 +0200 (Mon, 07 Jun 2010) | 1 line
Fixed a bug that causes Nping to fail when scanning an IP address assigned to a local network interface. The problem is complex and difficult to solve due to the way Nping handles targets and determines things like src IP addresses. This patch introduces an ugly hack (an explicit check for a particular situation in route_dst() ). This is not the most elegant solution and it probably only works in Linux, but it should be OK until a better solution is found.
------------------------------------------------------------------------
r17879 | luis | 2010-06-07 18:46:44 +0200 (Mon, 07 Jun 2010) | 1 line
Added a few wrappers for the new parse_unsigned_number() function. This should minimize the risks of calling it with an invalid pointer type. Also, its prototype has been removed from utils.h and it has been declared static. Additionally, the validate_number_spec() function was removed, relying on strtoul() to perform the appropriate validation.
------------------------------------------------------------------------
flag) is received in response. This can be due to an extremely rare TCP feature
known as a simultaneous open or split handshake connection.
see http://nmap.org/misc/split-handshake.pdf
Added a reason code: ER_SYN and associated reason string: "split-handshake-syn".
one already present in the current group. This is because ultra_scan
does not cope with multiple targets sharing the same IP address. However
this alone isn't enough to force a new host group, because the loop in
nmap.cc concatenates groups that nexthost splits apart, with its own
duplicate logic for breaking up groups.
different host group. Also insert the target into the group only after
we're sure we want it, instead of provisionally inserting it and then
removing it after we find it's unwanted.
We use this file to get at the LPADAPTER member of struct pcap on Windows,
which we need in order to pass to PacketSetReadTimeout in a few places.
(Normally pcap_t is an opaque type so we're not supposed to see inside it.) I
don't see an easy way to get rid of the calls to PacketSetReadTimeout right
now.
