Based on a tarball from Shell Hung <i@shellhung.org>
--
This module sprung out of a need to do one thing and one thing only,
do it securely, and do it well. This module creates and checks
detached signatures for data. That's it. If you want to do anything
else that PGP lets you do, look elsewhere.
--
Compared to the previous release, this version brings amongst other
changes the following:
o Support for giving interfacenames as internal/external address.
o contrib/ directory added.
o contrib/sockd-stat.awk, provides statistics based on sockd logfiles.
Contributed by Stephan Eisvogel <eisvogel@hawo.stw.uni-erlangen.de>.
o If gethostbyname() fails, treat it as if resolveprotocol was set to
fake, meaning we hope the socksserver will be able to resolve it.
Will presumably make certain dns configurations work better for
client.
See the NEWS file for a more complete list.
is to provide a powerful and extensible environment for solving
classical (pen-and-paper) ciphers, providing as much automation as
possible. Classical ciphers include common schemes like monoalphabetic
substitutions, where each letter of the alphabet is mapped to another
(usually different) letter consistently through the text. The first
version of Crank is restricting itself to these special ciphers. Other
algorithms forever devoid of Crank's attentions include Enigma, RSA,
DES, MurkelFish, or anything else invented after 1900.
NEWS:
- Due to an endianness handling problem Blowfish algorithm was not compatible
with other implementations. Now it has been corrected. If you want
to access the old algorithm used use the "blowfish-compat" module.
- Fixes in mcrypt_list_algorithms() for some systems. Bugs pointed out by
Jonathan Woolmington <jwool@ind.tansu.com.au>
- Fixes in stream mode.
- mcrypt_generic_init() no longer fails if smaller key is used. It uses
the most appropriate key size of the algorithm and pads with zeros.
- Fixes in wake algorithm (and support for IV).
- IV is now used in arcfour (arcfour-iv is now longer used).
Speedups in Arcfour.
disk all the files shared and the documents printed in a LanManager
environnement (all the Microsoft and Samba machines using LanManager
protocol to share data).
This is a proof of concept to show that LanManager (CIFS) is an
extremely insecure protocol.
It has been pointed out that there is another bug in the signature
verification code of GnuPG.
* This can easily lead to false positives *
All versions of GnuPG released before today are vulnerable!
To check a detached singature you normally do this:
gpg --verify foo.sig foo.txt
The problem here is that someone may replace foo.sig with a standard
signature containing some arbitrary signed text and its signature,
and then modify foo.txt - GnuPG does not detect this - Ooops.
The solution for this problem ist not easy and needs a change in the
semantics of the --verify command: It will not any longer be
possible to do this:
gpg --verify foo.sig <foo.txt
Instead you have to use this
gpg --verify foo.sig - <foo.txt
The difference here is that gpg sees 2 files on the command lines
and thereby knows that it should check a detached signature. We
really need this information and there is no way to avoid that
change, sorry. You should make sure that you never use the first
form, because this will lead to false positives when foo.sig is not
a detached signature - gnupg does detect the other case and warns
you, but this is not sufficient. If you use GnuPG from other
applications, please change it.
ok markus@
Version 3.9, 2000.12.13:
* Updated temporary key generation:
- stunnel is now honoring requested key-lengths correctly,
- temporary key is changed every hour.
* transfer() no longer hangs on some platforms.
Special thanks to Peter Wagemans for the patch.
* Potential security problem with syslog() call fixed.
--
mcrypt is intended to be a replacement of the old unix crypt(1)
under the GNU General Public License. Unix crypt(1) was a popular
file encryption program in unix boxes.
It was based on the Enigma encryption algorithm but it was considerable
trivialized. Since this was not adequate, even for individual privacy
needs, mcrypt was created as a similar program using some modern
block encryption algorithms.
Mcrypt also has a compatibility mode with unix crypt(1) and with
Solaris des(1). It supports all the algorithms and modes found in
libmcrypt and it is very extendable.
At the time writing this, it supports the algorithms: BLOWFISH,
TWOFISH, DES, TripleDES, 3-WAY, SAFER, LOKI97, GOST, RC2, RC6, MARS,
IDEA, RIJNDAEL, SERPENT, CAST, ARCFOUR and WAKE.
Block algorithms are implemented in modes: CFB, CBC, ECB, OFB (8
bit and n bit, where n is the size of the algorithm's block length).
For a brief description of the algorithms and the modes look at the
mcrypt manpage (this may be out of date). In mcrypt it is on the
user to decide which algorithm he considers best for encrypting his
data.
--
libmcrypt is the library which implements all the algorithms and
modes found in mcrypt. It is currently under development but it
seems to work pretty good.
Unlike most encryption libraries libmcrypt does not have everything
(random number generators, hashes, hmac implementation, key exchange,
public key encryption etc.). Libmcrypt only implements an interface
to access block and stream encryption algorithms.
Its purpose was to assist in the development of mcrypt by providing
a uniform interface to access several different encryption algorithms,
so that the main program is independent of the encryption algorithms
and the modes used.
Libmcrypt supports the algorithms: BLOWFISH, TWOFISH, DES, TripleDES,
3-WAY, SAFER-sk64, SAFER-sk128, SAFER+, LOKI97, GOST, RC2, RC6,
MARS, IDEA, RIJNDAEL-128 (AES), RIJNDAEL-192, RIJNDAEL-256, SERPENT,
CAST-128 (known as CAST5), CAST-256, ARCFOUR and WAKE. Block
algorithms can be used in: CBC, ECB, CFB and OFB (8 bit and n bit,
where n is the size of the algorithm's block length).
*** Tuesday, November 21, 2000 -- Dante v1.1.6
o fix a bug related to hostnamelength parsing in server.
Thanks to "Thomas Jarosch" <thomas.jarosch@styletec.de>.
*** Monday, October 16, 2000 -- Dante v1.1.5
o New prototype for gethostbyaddr in RedHat 7.0 added.
First reported by Paul R Streitman <prs@us.ibm.com>.
o RedHat needs libnsl for tcpwrappers to work.
*** Thursday, October 5, 2000 -- Dante v1.1.4
o fix bug affecting clients going through socks v4 servers.
Reported and nicely diagnosed by Jack Keane (jkeane@OpenReach.com).
o increase default listen backlog to 511, based on request by
Doug Hardie (bc979@lafn.org).
*** Monday, September 25, 2000 -- Dante v1.1.3
o some fixes/additions to example/ files.
o HP-UX 11.00 should now work.
Thanks to Malte Cornils <malte@cornils.net> for testing.
o httpproxysupport in client (meaning "socksify" can work
when going through webproxies too).
o expire badmarking on bad/non-working routes/proxyservers after
configured time. Default to never expiring, as in previous
versions. See BADROUTE_EXPIRE in config.h.
o say what address we expected the bindreply to come from in
"unexpected bindreply ..."
o don't close controlconnection if another socket is using it.
Fixes a bug triggered when using the bindextension in certain
cases. Problem reported by Jacques A. Vidrine (n@nectar.com).
o compilation outside source directory fixed, based on patch from
NISHIMURA Daisuke <nishi@graco.c.u-tokyo.ac.jp>
o bsdi uses elf; NISHIMURA Daisuke <nishi@graco.c.u-tokyo.ac.jp>
o dlib/hostcache.c now compiled again. First reported by
"Jacques A. Vidrine" <n@nectar.com>
--
This Perl module provides support for the https protocol under LWP,
so that a LWP::UserAgent can make https GET & HEAD & POST requests.
Please see 'perldoc LWP' for more information on POST requests.
- Decapitalize first letter of comment if appropriate.
- Remove trailing blank lines.
- Remove punctuation.
- Remove version numbers which are often overlooked when updating.
- espie@ ok
It now FAKEs, and installs in a nice clean CPAN way
- CONFIGURE_STYLE is perl now, remove a lot of old manual code
- remove all old patches, and add in a new one which cleans up
the program a bit (remove small linux specific hacks)
thanks to brad@ for the review
6 Sep 2000 (when RSA Security Inc. placed its public key encryption
algorithm into the public domain). Remove configure patch that
added the no-longer-needed --with-rsaref option
--
This patch prevents the client for sending zero length username/passwords
and prevents the server from accepting zero length username/passwords.
Only useful for sites using username/password authentication (rfc1929).
--
The patch is for clients that do not have DNS resolution. Sites that do
not use the remote name resolution do not have to apply this patch.
The fakehost files (/tmp/.s5fakehost-<uid>) created with releases prior to
release 10 are not compatible.
If the fakehost file has size 65284 bytes, it is the old version. Delete
the file, apply the following patch, and rebuild the clients. The new
fakehost file has size 32644 bytes.
- clean up Makefile and wrapper scripts some more
- better INSTALL/DEINSTALL script
- be able to compile w/ old nessus pkg installed (thx rdump@river.com)
- make it think nmap is always present to avoid dependency (espie@)
- fix backward handling of USA_RESIDENT vs. FLAVOR.
Flavored ports *MUST* set FLAVOR consistently, for the magic setting
of work, pkgname to work, e.g., any old port with variables such as
USA_RESIDENT must:
* set FLAVOR depending on those old option variables,
* only depend on FLAVOR for all subsequent tests.
and apply local patch instead of using the rsaref port directory.
Fixes problems trying to make both ports concurrently, among other
things.
This is a bookkeeping change only.. there is no change in the
generated libraries.
libssl/Makefile.bsd-wrapper did the wrong thing with make clean
if run before make prereq. bsd.port.mk doesn't clean up the obj
dir of the sslusa port when symlinked somewhere else. These two
bugs caused no end of grief to everyone but me. This is a workaround
for some of those bugs
- make with NOMAN set (this doesn't need to override system man pages)
- use a lndir instead of cp since patch makes new files and deals with
that ok
- by the above, take care of not accidentally bringing over obj dirs from a
build tree
