Home page logo

nmap-dev logo Nmap Development mailing list archives

Nmap 4.62 released
From: Fyodor <fyodor () insecure org>
Date: Sat, 3 May 2008 02:34:09 -0700

Hi all,

I'm pleased to release Nmap 4.62, which includes many great changes
made over the last month and a half.  This release has more than 3
dozen changes(!).  One of the most exciting is David's fixed-rate
packet sending system, which is something I've wanted in Nmap for
years.  Davids Mac OS X packaging work is also continuing and showing
great results.  And don't forget Brandon's version detection DB
improvements, Jah's fixing of the COPYING file display on Windows (a
long-time blight on that installer), and Kris's NSE Datafiles library.
There are many bug fixes and other improvements as well.

This is a nice pre-SoC stable Nmap.  Remember that we've got 7
full-time SoC coders starting in just a couple of weeks!  And most of
them have already started in a more limited capacity during this
"community bonding period".

I've linked to the new version from the Nmap download page:


Please give it a try and let us know on nmap-dev if you encounter any

Here is the (long!) list of changes since 4.60:

o Added a new --min-rate option that allows specifying a minimum rate
  at which to send packets. This allows you to override Nmap's
  congestion control algorithms and request that Nmap try to keep at
  least the rate you specify.  The rate is given in packets per
  second. Read more in the Nmap man page
  (http://nmap.org/book/man-performance.html) [David]

o Create /nmap/macosx directory in SVN with files necessary to build
  binary Mac OS X Nmap/Zenmap packages.  We are trying to create
  binary installer packages which are as useful and easy to use as the
  Windows installer.  This has involved a lot of work by David.  We
  aren't quite yet distributing the results on the Nmap download page,
  but testing our beta versions is useful.  You can find the latest
  universal (PPC and Intel) binary test version by looking at David
  Fifield's posts at http://seclists.org/nmap-dev/2008/q2/author.html .
  You can also read /nmap/macosx/README in svn for more info.

o Nmap 2008 Summer of Code students have began working (though full
  time doesn't start until late May).  Learn about the winners and their
  projects at http://seclists.org/nmap-dev/2008/q2/0132.html .

o Brandon added/modified a whole bunch of version detection signatures
  based on systems discovered when scanning UCSD's network.

o Reformat Nmap COPYING file (e.g. remove C comment markers, reduce
  line length) during Nmap windows build so that it looks much better
  when presented by the Windows executable (NSIS) installer.  Thanks
  to Jah for the patch, which was modified slightly by Fyodor.

o Added NSE Datafiles library which reads and parses Nmap's nmap-*
  data files for scripts.  The functions (parse_protocols(),
  parse_rpc() and parse_services()) return tables with numbers
  (e.g. port numbers) indexing names (e.g. service names).  The
  rpcinfo.nse script was also updated to use this library. [Kris]

o Fixed a bug in the nbase random number generator (and the way it
  interacted with Nmap and MS Windows) which caused duplicates in some
  instances.  Thanks to Jah for reporting the problem and working with
  Brandon Enright, Fyodor and Kris to fix it.

o It turns out that hours have 60 minutes, not 24.  Fixed a scan
  status message which was rolling over the hours column
  prematurely. [David]

o Added scripting options to Zenmap profile editor and command wizard
  to make use of NSE. [David]

o Zenmap now prints an exception message rather than segfaulting when it
  can't open a display (such as when trying to connect to an X server as
  an unauthorized user). Thanks to Aaron Leininger for the initial
  report and Guilherme Polo for suggesting the fix.

o Now ports in the "unfiltered" state can be selected for attention by
  NSE scripts. [Kris]

o Nbase random number generation system now avoids having a high-bit
  of zero in every other byte on Windows due to Windows having such a
  low RAND_MAX. [Jah]

o Added release dates for each Nmap version to this CHANGELOG going
  back to Nmap 3.00 (July 31, 2002).  Dates are in MM/DD/YY format.
  If someone wants to track down dates for the last 22% of the file
  (pre-3.00), you are welcome to do so and send a patch.  Searching
  Google for the version number and site:seclists.org seems to work
  well. [Fyodor]

o Nmap RPM builds now use the versions of libdnet, libpcap, libpcre,
  and liblua rather than whatever happens to be installed on the build
  system. [David]

o Zenmap can now be installed in and run in directories with a space
  in the name. [David]

o Fixed an assertion failure ("Target.cc:396: void
  Target::stopTimeOutClock(const timeval*): Assertion
  'htn.toclock_running == true' failed.")caused when a host had NSE
  scripts in multiple runlevels.  This also fixes --host-timeout
  behavior in NSE. [Kris]

o Reduce the maximum number of socket descriptors which Nmap is
  allowed to open concurrently.  This resoles a bug which could cause
  "Too many open files" error on Mac OS X when not running as
  root. [David]

o Canonicalized service names between nmap-service-probes (version
  detection DB) and nmap-services (port scanning DB). [Kris]

o Removed the "class" attribute from the tcpsequence element in XML
  output. For a long time it had always been "unknown class" because
  Nmap doesn't calculate a class anymore. The XML output version has
  been increased from 1.01 to 1.02. [David]

o Fixed a bug on Win32 which caused an infinite loop when Nmap
  encountered certain broadcast addresses. [Dudi Itzhakov]

o Fix MingW compilation by adding a signal.h include to
  main.cc. [Gisle Vanem]

o Fix the test in our build system to determine if liblua is already
  available or not. For example, the test needed to link with -lm
  since some systems require that.  [David].

o Added TIMEVAL_BEFORE and TIMEVAL_AFTER macros to test whether one
  timeval is earlier than another while avoiding possible integer
  overflows in a naive approach we were using previously. [David]

o Adjusted a bunch of code to avoid compilation warning messages on
  some Linux machines. [Andrew J. Bennieston]

o Fixed the NmapArpCache so that it actually works. Previously, Nmap
  was always falling back to the system ARP cache. Of course this
  raises the question of whether NmapArpCache is needed in the first
  place. [Daniel Roethlisberger]

o Fix a Zenmap bug which could cause the error message
  "zenmapCore.NmapOptions.OptionNotFound: No option named '' found!"
  if you create a new profile without checking any options then try to
  edit it. [David]

o Zenmap now shows a more helpful error message when there is an error
  in executing Nmap. [David]

o Zenmap now creates the directory ~/.zenmap-etc to store
  automatically generated GTK+ and Pango files. They used to go in the
  application bundle but that doesn't work on a read-only filesystem or
  disk image. This is what Wireshark does (~/.wireshark-etc), although
  the directory could be called anything. It doesn't have to persist
  across sessions.

o Added a mechanism in Zenmap for including extra executable search
  paths on specific platforms, so we can include /usr/local/bin in
  PATH on Mac OS X by default and add the Nmap install directory on
  Windows. [David]

o We now use --no-strip when building Zenmap Mac OS X packages to
  prevent many mysterious warnings which occur when the binary is
  stripped. [David]

o When Zenmap invokes Nmap, it now copies the whole environment for
  the Nmap invocation rather than just providing $PATH.  Windows may
  need this to do proper name resolution. [David]

o Corrected uptime parsing and reporting in SNMPsysdesr.nse for an
  uptime of less than 46 hours. [Kris]

o Modified the use of CXXFLAGS, CFLAGS, and CPPFLAGS in Nmap build
  system to work better when building Mac OS X universal
  binaries. [David]

o Added many additional PCRE option flags to the list returned by the
  NSE pcre.flags() function. [Kris]

o Changed the NSE function nmap.set_port_state() so that it checks to
  see if the requested port is already in the requested state.  This
  prevents "Duplicate port" messages during the script scan and the
  inaccurate "script-set" state reason. [Kris]

o Canonicalize NSE script license text--more than half did not even
  spell license correctly. They all still say that they are under Nmap's
  license, just with consistent capitalization and spelling, and now a
  link to Nmap legal page at http://nmap.org/man/man-legal.html.

o Updated ripeQuery.nse to not print extraneous whitespace. [Kris]

o Switched telnet brute force password cracking NSE (bruteTelnet.nse)
  to vulnerability category so it isn't executed by default.  It can
  take too long to run. [Eddie]

o NSE status messages now print host name and IP, rather than just the
  host name (which was blank when Nmap didn't know it). [Jah]

o Allocate 128 characters for the idle scan ScanProgressMeter title. Previously
  it was 32 characters. The "idle scan against " and the \0 terminator take up 19
  characters, leaving only 13, which isn't enough to represent all IP addresses,
  let alone host names. Bug reported by Stephan Fijneman, fixed by David.


Sent through the nmap-dev mailing list
Archived at http://SecLists.Org

  By Date           By Thread  

Current thread:
  • Nmap 4.62 released Fyodor (May 03)
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]