It is in pure python to avoid portability issues, since most DES
implementations are programmed in C (for performance reasons).
Triple DES class is also implemented, utilising the DES base. Triple DES
is either DES-EDE3 with a 24 byte key, or DES-EDE2 with a 16 byte key.
See the "About triple DES" section below more info on this algorithm.
The code below is not written for speed or performance, so not for those
needing a fast des implementation, but rather a handy portable solution
ideal for small usage.
Upstream changes:
2.31 Tue Oct 30 07:03:40 EDT 2012
- Fixes to regular expressions to avoid rare failures to
correctly strip padding in decoded messages.
- Add padding type = "none".
- Both fixes contributed by Bas van Sisseren.
Today you need to remember many passwords. You need a password for
the Windows network logon, your e-mail account, your website's FTP
password, online passwords (like website member account), etc. etc.
etc. The list is endless. Also, you should use different passwords
for each account. Because if you use only one password everywhere
and someone gets this password you have a problem... A serious
problem. The thief would have access to your e-mail account, website,
etc. Unimaginable.
KeePass is a free open source password manager, which helps you to
manage your passwords in a secure way. You can put all your passwords
in one database, which is locked with one master key or a key file.
So you only have to remember one single master password or select
the key file to unlock the whole database. The databases are
encrypted using the best and most secure encryption algorithms
currently known (AES and Twofish).
ChangeLog Since 2.3.5:
* fix a bug when receiving a signature using the InclusiveNamespaces
PrefixList by copying namespace declaration from upper level at the level of
the signed node.
* fix compilation warning on recent version of GCC
1.49 2012-09-25
Fixed problem where on some platforms test t/local/07_tcpecho.t would
bail out if it could not bind port 1212. Now now tries a number of ports to bind to until
successful.
Improvements to unsigned casting contributed by Reini Urban.
Improvements to Net::SSLeay::read to make it easier to use with non-blocking IO:
contributed by James Marshall: It modifies
Net::SSLeay::read() to return the result from SSL_read() as the second
return value, if Net::SSLeay::read() is called in list context. Its
behavior should be unchanged if called in scalar or void context. This
result code seems to be required for full support of non-blocking I/O,
since users need to handle SSL_ERR_WANT_READ, SSL_ERROR_WANT_WRITE, etc.
Fixed a problem where t/local/kwalitee.t fails with
Module::CPANTS::Analyse 0.86. Patch from Paul.
Fixed a number of typos patched by Giles.
Fixed a compiler warning from Compiling with gcc-4.4 and -Wall, patched by Giles.
Fixed problems with get_https4: documentation was wrong, $header_ref was
not correctly set and $server_cert was not returned.
Fixed a problem that could cause a Perl exception about no blength
method on undef. Reported by "Stephen J. Smith via RT". https://rt.cpan.org/Ticket/Display.html?id=79309
Added documentation about how to mitigatxe various SSL/TLS
vulnerabilities.
Fixed problem reported by Mike Doherty: SSL_MODE_* are defined in ssl.h,
and should be available as constants, but I do not see them listed in constants.h
5.72 Mon Sep 24 15:22:08 MST 2012
- adjusted module installation directory for later Perls
-- As of 5.11 Perl searches 'site' first, so use that
-- ref. INSTALLDIRS in Makefile.PL
-- thanks to Robert Sedlacek for patch
0.64 2012-08-06 01:23:30
- Drastically simplify Makefile.PL to resolve RT bugs #61249, #61324,
#63553, #68208, and #68084.
- Forgot to update Changes for 0.63, so this version overrides that.
0.61_05 2012-08-04 00:40:22 UTC
- Trying to distinguish between good vs bad zero returns from underlying
SSL_read/SSL_write broke stuff (see RT bug #78695). Revert to previous
behavior.
- Completely re-organize Makefile.PL. I hope these changes will help take
care of RT bugs #61324, #61249, #63553, and #68084 etc. This is not
necessarily finished, but I want to see what happens on CPAN Testers at
this point before making a few other minor changes.
0.60 2012-07-29 21:43:47 UTC
- Release 0.59_03 as 0.60 so distributions can pick up various fixes. The most
important one seems to be bug RT #70565. This should take care of bug RT
#77167
- SSL_read and SSL_write now try to handle incomplete reads/writes (see bug RT
RT #64054). The current test suite does is not very comprehensive, so caution
is recommended at this point. Also, if you have good test cases, I would love
to incorporate them into the distribution.
0.59_03 2012-03-10 00:45:28 UTC
- Bump version number and upload to CPAN.
0.59_02 2012-03-08 16:16:03 UTC
- Forgot to update Changes for 0.59_01. The following is a combined list of
the more important fixes incorporated in both.
- Bug RT #64054: Handle incomplete reads/writes better
- Bug RT #73754: Add LWP::Protocol::https to PREREQ_PM
- Bug RT #73755: Crypt-SSLeay does not verify hosts (yet). Don't let that
cause a failure during tests.
- Streamline t/02-live.t using Try::Tiny and done_testing
- Plus assorted related small changes.
0.58_01 2010-09-08 19:11:39 UTC
- L<text|scheme:...> is not supported in POD for 5.8.5 and earlier.
- TODOs in POD should stand out
- Add /boot/common/ssl and some other directories to unix_ssl_dirs (see bug
#60936).
of the scrypt key derivation function. On modern hardware and with default
parameters, the cost of cracking the password on a file encrypted by scrypt
enc is approximately 100 billion times more than the cost of cracking the
same password on a file encrypted by openssl enc; this means that a five-
character password using scrypt is stronger than a ten-character password
using openssl
Generated file didn't pass -Werror check on gcc4.7. The problem is
well-known and already fixed on the current version of mit-krb5. The
patch added here was taken from upstream.
No revbump necessary, won't change binary on systems that already built it.
* 1.8.14, 2012-07-18
- The malloc allocator would return null instead of throwing in the event of
an allocation failure, which could cause an application crash due to null
pointer dereference where normally an exception would occur.
- Recent versions of OpenSSL include extra information in ECC private keys,
the presence of which caused an exception when such a key was loaded by
botan. The decoding of ECC private keys has been changed to ignore these
fields if they are set.
- AutoSeeded_RNG has been changed to prefer /dev/random over /dev/urandom
- Fix detection of s390x (Debian bug 638347)
In the Makefile, remove PYDISTUTILSPKG=yes, add
.include "../../lang/python/egg.mk", and set PKGREVISION=1.
This fixes errors from "make package" complaining about files
that were installed but not present in the PLIST.
Support for TLSA records.
Support for ILNP (NID, L64, L3, LP) records (untested).
Support for IPSECKEY records.
Handle TYPEXXX for known types correctly.
A number of NSEC3-related bug fixes.
Miscellaneous bug fixes.
* Noteworthy changes in release 2.14 (2012-09-24) [stable]
- Added asn1_read_node_value() to obtain a node's value.
This is to deprecate the export of the node_asn internal structure
for the upcoming 3.x release. The ASN1_DATA_NODE type and the
ASN1_ETYPE_* constants were added to support the new function.
2011-03-06 Ludovic Rousseau <rousseau@debian>
* README: release 1.4.12
* PCSC.xs: Also check for SVt_PVIV and not just SVt_IV.
Fixes Debian bug #613722 "libpcsc-perl: GetStatusChange error after print"
* PCSC.xs, PCSC.pm, test.pl: Use lines of less than 80 columns
2010-10-27 Ludovic Rousseau <rousseau@debian>
* create_distrib.sh: compress using bzip2 instead of gzip
* README: release 1.4.11
* PCSCperl.h: SCardCancelTransaction() is no more present in pcsc-lite
* PCSC.pm, PCSC.pod, PCSC.xs, PCSCperl.h:
SCardSetTimout is no more present in pcsc-lite > 1.6.4
It was a pcsc-lite specific and deprecated function
2010-08-18 Ludovic Rousseau <rousseau@debian>
* PCSCperl.h:
use SCARD_READERSTATE * instead of LPSCARD_READERSTATE since is not
define in pcsc-lite < 1.6.3
* README: release 1.4.10
* PCSC.pm, PCSC.xs, PCSCperl.h, README: Update copyright date
* create_distrib.sh: Allow more than one digit in the version numbers
* PCSC.xs, PCSCperl.h:
Use LPSCARD_READERSTATE instead of LPSCARD_READERSTATE_A since it is no
more defined in pcsc-lite >= 1.6.2
2010-06-30 Ludovic Rousseau <rousseau@debian>
* README: release 1.4.9
* PCSC.pm, PCSC.pod, PCSC.xs: sort constants in alphabetical order
* PCSC.pod, PCSC.xs:
SCARD_E_UNSUPPORTED_FEATURE is not specific to pcsc-lite but is also
available on Windows
* PCSC.pm, PCSC.pod, PCSC.xs:
SCARD_W_INSERTED_CARD is no more defined in pcsc-lite 1.6.0 and then
pcsc-perl failed to build. It was an error code specific to pcsc-lite.
2010-01-03 Ludovic Rousseau <rousseau@debian>
* Card/Card.pod: fix spelling error
2009-09-23 Ludovic Rousseau <rousseau@debian>
* README: release 1.4.8
* Card/Card.pod: Correct 3 bugs signaled by podchecker
2009-09-06 Ludovic Rousseau <rousseau@debian>
* PCSC.xs:
PCSC.xs:853: warning: format ‘%d’ expects type ‘int’, but argument 2 has type ‘long unsigned int’
2008-09-28 Ludovic Rousseau <rousseau@debian>
* PCSC.pod: example code for GetStatusChange()
=== 0.4.7 2012-09-03
* Fix merging paths if the path is not empty
* Set a configurable timeout for all requests (Rick Olson)
* Fix nested hash params in Consumer#request (Ernie Miller)
The PLIST was hardcoded for libtls150.so. Build With tcl 8.5, the
generated library becomes libtls85.so. Use the tcl Makefile.version
file to generate a PLIST_SUBST variable to make this future-proof.
This is a bugfix release.
* Fix KDC uninitialized pointer vulnerabilities that could lead to a denial of
service [CVE-2012-1014] or remote code execution [CVE-2012-1015].
* Correctly use default_tgs_enctypes instead of default_tkt_enctypes for TGS
requests.
This update of ocaml-cryptokit to its newest version, 1.6. does
not actually change anything in the functionality of the software.
Upstream, the build system has changed, and there are also a few
changes to the package (most notably, removal of the PLIST.opt in
favour of the PLIST_VARS system)
changes: fix a use-after-free bug which could be used to potentially
execute arbitrary code with root privileges, provided that the user
has been authenticated using a public key and also that a command
restriction is enforced (the "command" option must be used in
the authorized_keys file)
Bugfixes:
* SUPPORT-30: RRSIGs are left in the signed zone when authoritative RRsets
become glue [OPENDNSSEC-282].
* OPENDNSSEC-261: Ldns fails to parse RR that seems syntactically correct.
Was due to memory allocation issues. Provided better log message.
* OPENDNSSEC-285: Signer segfault for 6 or more -v options
* OPENDNSSEC-298: Only unlink existing pidfile on exit if we wrote it.
* OPENDNSSEC-303: Return if open/parse of zonelist.xml fails in ksmutil.c
update_zones() and cmd_listzone().
* OPENDNSSEC-304: Signer Engine: Check pidfile on startup, if pidfile exists
and corresponding process is running, then complain and exit.
* Signer seems to hang on a ods-signer command. Shutdown client explicitly
with shutdown().
* opendnssec.spec file removed
Bison 2.6.x+ handles the yydebug functionality differently by predefining
YYDEBUG. The yacc logic is not expecting YYDEBUG to be defined without
an value, so it breaks on an "#if YYDEBUG" macro in a few places.
In order to make this work with pre-2.6 bison as well as current versions,
hard code the inclusion of debug symbols. It doesn't hurt anything and
it fixes the package.
dhbitty is a small public key encryption program written in C. It
uses elliptic curve Diffie-Hellman in the form of Curve25519 to
establish a shared secret between two users, and uses that secret to
symmetrically encrypt and authenticate messages.
There are no private key files; only passphrases. Never lose that
pesky thing again.
Both the sender and the receiver can decrypt a message. In fact,
there is no distinction between sender and receiver. Both passphrases
must be strong.
There is no signing. A similarly useful form of authentication occurs
using only DH. dhbitty attempts to be as simple as possible. It is
not optimized, but achieves a comfortable speed for most uses. It
does not use floating point numbers, or integers longer than 32 bits.
It does not contain more algorithms than are needed.
Example
This is how Alice generates her public key with dhbitty:
$ dhbitty generate alice_public_key.txt
username:passphrase (this is visible!): alice:Keyfiles be damned!
Done.
Bob will do the same thing:
$ dhbitty generate bob_public_key.txt
username:passphrase (this is visible!): bob:Bob's Spectacular Passphrase
Done.
Alice will publish her alice_public_key.txt, and Bob will publish his bob_public_key.txt. They can now access each other's
public keys. (But they should be careful that Eve cannot surreptitiously replace either public key with her own!)
Alice wants to send files to Bob. She packages them into a .tar archive (or any other type of archive with timestamps), along
with her message. Then she uses dhbitty:
$ dhbitty encrypt bob_public_key.txt files_to_bob.tar files_to_bob.tar.dhbt
username:passphrase (this is visible!): alice:Keyfiles be damned!
Done.
Alice sends files_to_bob.tar.dhbt to Bob. Bob will use dhbitty to decrypt this archive:
$ dhbitty decrypt files_to_bob.tar.dhbt files_to_bob.tar
username:passphrase (this is visible!): bob:Bob's Spectacular Passphrase
This is the public key of file's secondary owner:
0002f02b318c307bac07f3148a33c975cea04b79a870f0a5c7771cd38cc1986e
Done.
Bob can verify that the public key dhbitty just gave him indeed is Alice's public key. He unpacks the now-decrypted archive to
access the files Alice sent to him.
In practice, Alice and Bob should use a system like diceware to pick passphrases, in order to be confident of their strength.
Seven words picked using diceware is a good choice.
Note: Nobody that uses git from pkgsrc can install this package.
It conflicts with security/heimdal which is sucked in by dependencies
of scmgit-base. Since the default way of acquiring pkgsrc on
DragonFly is via git, which is provided by the releases and daily
snapshots, effectively this can't be installed by DragonFly users.
Solving the conflict with heimdal, if possible, would be nice.
When this package was updated to version 1.1, it stopped building on
DragonFly. The main issue is that DragonFly doesn't have bind in its
base. NetBSD does, so it zkt finds it there, but otherwise it needs
the configuration switch --enable-binutil-path to be used. This was
added for DragonFly to point at ${PREFIX}/sbin.
zkt requires bind to be installed in order to build. Unlike other
packages like python, postgresql, and ruby where the mk.conf can
define a version otherwise a default is used, no such mechanism
exists to hand the four separate bind packages (at least I don't know
about one). So the inclusion of bind99 is a hack I'm not too proud
of, but I don't have a better solution. With it, it builds in clean
environments like pbulk chroot and Tinderbox. If an individual user
is building from source, they'll be smart enough to comment out this
include if another version of bind is already installed (zkt will
fail on a bind build conflict).
I suspect DragonFly is rather unique in not having bind in base, so
for now this is left as a DragonFly-specific section. Something
like net/bind99/builtin.mk could possibly be used to determine if
no builtin bind is available and thus follow DragonFly approach. I
shall leave it to others to decide.
GNUTLS deprecated gnutils_transport_set_lowat function in version 2.12.0
and finally removed it with version 3.0, breaking any packages that
still reference it.
The lowat feature is now disabled permanently I think.
The patch uses the GNUTLS_VERSION_NUMBER macro to appropriately conceal
the function reference. The same patch is widely seen on the 'net with
other packages that use gnutls like OpenVAS.
is starttls's implementation is incompatible with emacs 22, 23 and probably
24 too, as a result sending emails with ssl/tls authorization fail due to
this
conflict. emacs-21 has its own starttls.el too and I believe it is also
sufficient. I wonder if someone still uses emacs-20 and its smtpmail.el for
sending emails. This change was tested on NetBSD-6 and emacs-{22,23}.
starttls package now DEPENDS on emacs-[0-9]*, that is any emacs flavour is
good enough
Set LICENSE to gnu-gpl-v2
++pkgrevision
It seems that I386 DragonFly (x86_64 is okay), invoking libintl's
bindtextdomain causes pkgsrc's libintl to segfault on a thread
locking operation. Anything linking with libgpg-error on i386
will consequently core dump.
Recognizing this treating the symptom, this patch disabled NLS on
I386 DragonFly.
This is a bugfix release.
* Fix an interop issue with Windows Server 2008 R2 Read-Only Domain Controllers.
* Update a workaround for a glibc bug that would cause DNS PTR queries to occur
even when rdns = false.
* Fix a kadmind denial of service issue (null pointer dereference), which could
only be triggered by an administrator with the "create" privilege.
[CVE-2012-1013]
Changes 1.10.1:
This is a bugfix release.
* Fix access controls for KDB string attributes [CVE-2012-1012]
* Make the ASN.1 encoding of key version numbers interoperate with Windows
Read-Only Domain Controllers
* Avoid generating spurious password expiry warnings in cases where the KDC
sends an account expiry time without a password expiry time.
0.4.6 (2011-10-16)
=====
* Added write_certificate function.
* Remove support for SSLv2, which was dropped upstream (thanks Dario Teixeira).
* Added support for compiling under Win32 (thanks David Allsopp), see
README.win32.
* Check for pthreads in configure.
0.4.5 (2011-03-01)
=====
* Use pthread mutexes for locking thread-safe version of ssl.
0.4.4 (2010-01-06)
=====
* Use SSL_CTX_use_certificate_chain_file instead of
SSL_CTX_use_certificate_file.
* Added support for --enable-debugging configure option.
* Don't link with unix library and don't build in custom mode.
0.4.3 (2008-12-18)
=====
* Don't use blocking sections in finalizers since it causes segfaults (thanks
Grégoire Henry and Stéphane Glondu).
from 2.51nb1 to 2.52.
Upstream changes:
2012-06-08 Gisle Aas <gisle@ActiveState.com>
Gisle Aas (3):
Wrong version number in the changelog
The t/threads.t was missing from the MANIFEST
Update expected digests for files
Andrew Fresh (1):
Remove double the
Lyle Hopkins (1):
Digest::Perl::MD5 OO fallback didn't work [RT#66634]
Peter J. Acklam (1):
Fix typos (spelling errors) in cpan/Digest-MD5/*
Shlomi Fish (1):
Modernize the code in the POD.
Zefram (1):
Makes Digest::MD5 work on Perl 5.6 [RT#75032]
security/p5-IO-Socket-SSL from 1.74 to 1.76.
Upstream changes:
v1.76 2012.06.18
- no longer depend on Socket.pm 1.95 for inet_pton, but use Socket6.pm if
no current Socket.pm is available. Thanks to paul[AT]city-fan[DOT]org
for pointing out the problem and providing first patch
v1.75 2012.06.15
- made it possible to explicitly disable TLSv11 and TLSv12 in SSL_version
changes: bugfixes:
-Fixed memory leak in PKCS #8 key import
-Check key identifiers when checking for an issuer
pkgsrc note: This is just a last checkpoint on the 2.x branch, in case
it will be needed for the Q2 branch. Will update to 3.x RSN.
* libclamav: Scan output at end of truncated tar
* libclamav: Fix handling of tar file with malformed header
* libclamav: Scan chm with invalid handling
* freshclam: give custom dbs higher priority during update
* libclamav: detect read races and abort the scan with an error
* libclamav/pe.c: drop old header check
Upstream changes:
-----------------
## ssh 1.7.14 (2012-05-07)
* #15: Implemented parameter substitution in SSHConfig, matching the
implementation of `ssh_config(5)`. Thanks to Olle Lundberg for the patch.
* #24: Switch some internal type checking to use `isinstance` to help prevent
problems with client libraries using subclasses of builtin types. Thanks to
Alex Morega for the patch.
* [Fabric #562](https://github.com/fabric/fabric/issues/562): Agent forwarding
would error out (with `Authentication response too long`) or freeze, when more
than one remote connection to the local agent was active at the same time.
This has been fixed. Thanks to Steven McDonald for assisting in
troubleshooting/patching, and to GitHub user `@lynxis` for providing the
final version of the patch.
===
F-PROT Antivirus for Unix, version 6.2.1
Compatibility for older Linux distros improved (glibc 2.3 for 32 bit version and glibc 2.4 for 64 bit version)
Compatibility for older Solaris/SunOS version improved (both 32 and 64 bit versions are compatible with solaris 8 now)
64 bit FreeBSD now supported
===
F-PROT Antivirus for Unix, version 6.2.0
Scan engine upgraded from 4.6.2 to 4.6.5 with improved detection rates and fewer false positives.
Multiple issues with the mail scanners have been fixed.
===
F-PROT Antivirus for Unix, version 6.1.1
fpupdate fix to prevent crash on certain 64 bit Linux systems.
* OPENDNSSEC-277: Enforcer: Performance optimisation of database access.
Bugfixes:
* SUPPORT-27: ods-ksmutil: simplify zone delete so that it only marks keys as
dead (rather than actually removing them). Leave the key removal to purge
jobs.
(Ok'ed by wiz@)
* misc Release numbering changed to three level "major.minor.revison" scheme
* bug REMOVE_HOLD_TIME was set to 10 days only (Thanks to Chris Thompson)
* doc Improved README file (Thanks to Jan-Piet Mens)
* misc Fix of some typos in log messages
* bug Fixed error in rollover.c (return code of genfirstkey() wasn't checked)
* misc Default of KeySetDir changed from NULL to ".." (best for hierarchical mode)
Default Sig Lifetime changed from 10 days to 3 weeks (21 days)
Default ZSK lifetime changed from 3 months to 4 times the sig lifetime
Default KSK lifetime changed from 1 year to 2 years
Parameter checks in checkconfig() adapted.
KSK random device changed back from /dev/urandom to BIND default
(Be aware of some possibly long delay in key generation)
* func New configure option to set the bind utility path manually (--enable-bindutil_path)
BIND_UTIL_PATH in config_zkt.h will no longer used
* bug If nsec3 is turned on and KeyAlgo (or AddKeyAlgo) is RSHASHA1
or DSA, genkey() uses algorithm type NSECRSASHA1 or NSEC3DSA instead.
* bug Error in printconfigdiff() fixed. (Thanks to Holger Wirtz)
* func Description added to (some of the) dnssec.conf parameters
* func Adding a patch from Hrant Dadivanyan to always pre-publish ZSKs
* misc Config file syntax changed to parameter names without underscores.
zkt-conf uses ZKT_VERSION string as config version
* bug "make install-man" now installs all man page
* bug Bug fixed in zfparse.c. zkt-conf was unable to detect an already
included dnskey.db file if another file was included.
* misc destination dnssec-zkt removed from Makefile.in
* func dki_prt_managedkeys() added to dki.c
zkt_list_managedkeys() added to zkt.c
zkt-ls has new option -M to print out a list of managed-keys
* bug Bug fixed in the config parser (zconf.c). Couldn't parse
agorithm RSASHA512 correctly (Thanks to Michael Sinatra)
* Add an anon_fast option that attempts anonymous authentication
(generally implemented via anonymous PKINIT inside the Kerberos
library) and then, if successful, uses those credentials for FAST
armor. If fast_ccache and anon_fast are both specified, anonymous
authentication will be used as a fallback if the specified FAST ticket
cache doesn't exist. Based on patches from Yair Yarom.
* Add a user_realm option to only set the realm for unqualified user
principals. This differs from the existing realm option in that realm
also changes the default realm for authorization decisions and for
verification of credentials. Update the realm option documentation to
clarify the differences and remove incorrect information. Patch from
Roland C. Dowdeswell.
* Add a no_prompt option to suppress the PAM module's prompt for the
user's password and defer all prompting to the Kerberos library. This
allows the Kerberos library to have complete control of the prompting
process, which may be desireable if authentication mechanisms other
than password are in use. Be aware that, with this option set, the
PAM module has no control over the contents of the prompt and cannot
store the user's password in the PAM data. Based on a patch by Yair
Yarom.
* Add a silent option to force the module to behave as if the
application had passed in PAM_SILENT and suppress text messages and
errors from the Kerberos library. Patch from Yair Yarom.
* Add preliminary support for Kerberos trace logging via a trace option
that enables trace logging if supported by the underlying Kerberos
library. The option takes as an argument the file name to which to
log trace output. This option does not yet work with any released
version of Kerberos, but may work with the next release of MIT
Kerberos.
* MIT Kerberos does not add a colon and space to its password prompts,
but Heimdal does. pam-krb5 previously unconditionally added a colon
and space, resulting in doubled colons with Heimdal. Work around this
inconsistency by not adding the colon and space if already present.
* Fix alt_auth_map support to preserve the realm of the authentication
identity when forming the alternate authentication principal, matching
the documentation.
* Document that the alt_auth_map format may contain a realm to force all
mapped principals to be in that realm. In that case, don't add the
realm of the authentication identity. Note that this can be used as a
simple way to attempt authentication in an alternate realm first and
then fall back to the local realm, although any complex attempt at
authentication in multiple realms should instead run the module
multiple times with different realm settings.
* Avoid a NULL pointer dereference if krb5_init_context fails.
* Fix initialization of time values in the module configuration on
platforms (like S/390X) where krb5_deltat is not equivalent to long.
* Close a memory leak when search_k5login is set but the user has no
.k5login file.
* Close several memory leaks in alt_auth_map support.
* Suppress bogus error messages about unknown option for the realm
option. The option was being parsed and honored despite the error.
* Retry authentication under try_first_pass on several other errors in
addition to decrypt integrity check errors to handle a wider array of
possible "password incorrect" error messages from the KDC.
* Update to rra-c-util 4.4:
* Update to C TAP Harness 1.12:
- Bux fix release
- Rollerd's -alwayssign flag logic had a critical error that could
have caused a zone to be signed with the wrong ZSK at particular
points of the ZSK key rolling process.
* Only use libyubikey when --with-cr is used.
* Set correct permissions on tempfile.
* YubiKey 2.2 contains a bug in challenge-response that makes it output the
same response to all challenges unless HMAC_LT64 is set. Add warnings to
ykpamcfg and a warning through conversate in the pam module. Keys programmed
like this should be reprogrammed with the HMAC_LT64 flag set.
* Implement option -ooath-id to easily set OATH token identifier.
* Fix numerous compiler warnings from clang. Thanks to
Clemens Lang <neverpanic@gmail.com>.
* ykclient: Add C++ namespace protection.
* Add multi-server support with curl_multi.
Enabled by default for YubiCloud servers.
Settable with the new library function set_template_urls() or
the urls parameter to ykclient_verify_otp_v2().
* Remove extra % in ykclient help.
* Add ca path option to ykclient, --ca.
Patch from Jay Kline <jay.kline.ctr@hpcmo.hpc.mil>.
* Make the nonce unique for consecutive calls to the same ykclient handle.
* Do url encoding of OTP before sending.
* Fix segfault on curl error.
Patch from Lee Hinman <lee.hinman.ctr@hpc.mil>
decentralized, and highly reliable synchronization. That means that a key
submitted to one SKS server will quickly be distributed to all key servers,
and even wildly out-of-date servers, or servers that experience spotty
connectivity, can fully synchronize with rest of the system.
