2.043 2017/01/06
- make t/session_ticket.t work with OpenSSL 1.1.0. With this version the
  session does not get reused any longer if it was not properly closed which
  is now done using an explicit close by the client which causes a
  proper SSL_shutdown
2.042 2017/01/05
- enable session ticket callback with Net::SSLeay>=1.80
2.041 2017/01/04
- leave session ticket callback off for now until the needed patch is
  included in Net::SSLeay. See
2.040 2016/12/17
- fix detection of default CA path for OpenSSL 1.1.x
- Utils::CERT_asHash now includes the signature algorithm used
- Utils::CERT_asHash can now deal with large serial numbers
2.039 2016/11/20
- OpenSSL 1.1.0c changed the behavior of SSL_read so that it now returns -1 on
  EOF without proper SSL shutdown. Since it looks like that this behavior will
  be kept at least for 1.1.1+ adapt to the changed API by treating errno=NOERR
2.038 2016/09/17
- restrict session ticket callback to Net::SSLeay 1.79+ since version before
  contains bug. Add test for session reuse
- extend SSL fingerprint to pubkey digest, i.e. 'sha1$pub$xxxxxx....'
- fix t/external/ocsp.t to use different server (under my control) to check 
  OCSP stapling
2.037 2016/08/22
- fix session cache del_session: it freed the session but did not properly
  remove it from the cache. Further reuse causes crash.
2.036 2016/08/11
- disable OCSP support when Net::SSLeay 1.75..1.77 is used, see RT#116795
2.035 2016/08/11
- fixes for issues introduced in 2.034
  - return with error in configure_SSL if context creation failed. This
    might otherwise result in an segmentation fault later.
  - apply builtin defaults before any (user configurable) global settings
    (i.e. done with set_defaults, set_default_context...) so that builtins
    don't replace user settings
    Thanks to joel[DOT]a[DOT]berger[AT]gmail[DOT]com for reporting
2.034 2016/08/08
- move handling of global SSL arguments into creation of context, so that these
  get also applied when creating a context only.
2.033 2016/07/15
- support for session ticket reuse over multiple contexts and processes
  (if supported by Net::SSLeay)
- small optimizations, like saving various Net::SSLeay constants into variables
  and access variables instead of calling the constant sub all the time
- make t/dhe.t work with openssl 1.1.0
2.032 2016/07/12
- Set session id context only on the server side. Even if the documentation for
  SSL_CTX_set_session_id_context makes clear that this function is server side
  only it actually affects hndling of session reuse on the client side too and
  can result in error "SSL3_GET_SERVER_HELLO:attempt to reuse session in
  different context" at the client.
2.031 2016/07/08
- fix for bug in session handling introduced in 2.031, RT#115975
  Thanks to paul[AT]city-fan[DOT]org for reporting
2.030 2016/07/08
- Utils::CERT_create - don't add given extensions again if they were already
  added. Firefox croaks with sec_error_extension_value_invalid if (specific?)
  extensions are given twice.
- assume that Net::SSLeay::P_PKCS12_load_file will return the CA certificates
  with the reverse order as in the PKCS12 file, because that's what it does.
- support for creating ECC keys in Utils once supported by Net::SSLeay
- remove internal sub session_cache and access cache directly (faster)
2.029 2016/06/26
- fix del_session method in case a single item was in the cache
- use SSL_session_key as the real key for the cache and not some derivate of it,
  so that it works to remove the entry using the same key
2.028 2016/06/26
- add del_session method to session cache
2.027 2016/04/20
- only added Changes for 2.026
2.026 2016/04/20
- update default server and client ciphers based on recommendation of
  Mozilla and what the current browsers use. Notably this finally disables
  RC4 for the client (was disabled for server long ago) and adds CHACHA20.
2.025 2016/04/04
- Resolved memleak if SSL_crl_file was used: RT#113257, RT#113530
  Thanks to avi[DOT]maslati[AT]forescout[DOT]com and
  mark[DOT]kurman[AT]gmail[DOT]com for reporting the problem
2.024 2016/02/06
- Work around issue where the connect fails on systems having only a loopback
  interface and where IO::Socket::IP is used as super class (default when
  available). Since IO::Socket::IP sets AI_ADDRCONFIG by default connect to
  localhost would fail on this systems. This happened at least for the tests,
  see https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813796
  Workaround is to explicitely set GetAddrInfoFlags to 0 if no GetAddrInfoFlags
  is set but the Family/Domain is given. In this case AI_ADDRCONFIG would not
  be useful anyway but would cause at most harm.
2.023 2016/01/30
- OpenSSL 1.0.2f changed the behavior of SSL shutdown in case the TLS connection
  was not fully established (commit: f73c737c7ac908c5d6407c419769123392a3b0a9).
  This somehow resulted in Net::SSLeay::shutdown returning 0 (i.e. keep trying)
  which caused an endless loop. It will now ignore this result in case the TLS
  connection was not yet established and consider the TLS connection closed
2.022 2015/12/10
- fix stringification of IPv6 inside subjectAltNames in Utils::CERT_asHash.
  Thanks to Mark.Martinec[AT]ijs[DOT]si for reporting in #110253
2.021 2015/12/02
- Fixes for documentation and typos thanks to DavsX and jwilk.
- Update PublicSuffx with latest version from publicsuffix.org
2.020 2015/09/20
- support multiple directories in SSL_ca_path as proposed in RT#106711
  by dr1027[AT]evocat[DOT]ne. Directories can be given as array or as string
  with a path separator, see documentation.
- typos fixed thanks to jwilk https://github.com/noxxi/p5-io-socket-ssl/pull/34
2.019 2015/09/01
- work around different behavior of getnameinfo from Socket and Socket6 by
  using a different wrapper depending on which module I use for IPv6.
  Thanks to bluhm for reporting.
2.018 2015/08/27
- RT#106687 - startssl.t failed on darwin with old openssl since server 
  requested client certificate but offered also anon ciphers
2.017 2015/08/24
- checks for readability of files/dirs for certificates and CA no longer use
  -r because this is not safe when ACLs are used. Thanks to BBYRD, RT#106295
- new method sock_certificate similar to peer_certificate based on idea of
  Paul Evans, RT#105733
- get_fingerprint can now take optional certificate as argument and compute
  the fingerprint of it. Useful in connection with sock_certificate.
- check for both EWOULDBLOCK and EAGAIN since these codes are different on
  some platforms. Thanks to Andy Grundman, RT#106573
- enforce default verification scheme if none was specified, i.e. no longer
  just warn but accept. If really no verification is wanted a scheme of 
  'none' must be explicitly specified.
- support different cipher suites per SNI hosts
2.016 2015/06/02
- add flag X509_V_FLAG_TRUSTED_FIRST by default if available in OpenSSL 
  (since 1.02) and available with Net::SSLeay. RT#104759 (thanks GAAS)
- work around hanging prompt() with older perl in Makefile.PL RT#104731
- make t/memleak_bad_handshake.t work on cygwin and other systems having 
  /proc/pid/statm, see RT#104659
- add better debugging based on patch from H.Merijn Brand
2.015 2015/05/13
- work around problem with IO::Socket::INET6 on windows, by explicitly using
  Domain AF_INET in the tests. 
  Fixes RT#104226 reported by CHORNY
2.014 2015/05/05
- Utils::CERT_create - work around problems with authorityInfoAccess, where
  OpenSSL i2v does not create the same string as v2i expects
- Intercept - don't clone some specific extensions which make only sense with
  the original certificate
2.013 2015/05/01
- assign severities to internal error handling and make sure that follow-up
  errors like "configuration failed" or "certificate verify error" don't
  replace more specific "hostname verification failed" when reporting in
  sub errstr/$SSL_ERROR. see also RT#103423
- enhanced documentation thanks to Chase Whitener 
2.012 2015/02/02
- fix t/ocsp.t in case no HTTP::Tiny is installed
2.011 2015/02/01
- fix t/ocsp.t - don't count on revoked.grc.com using OCSP stapling #101855
- added option 'purpose' to Utils::CERT_create to get better control of the
  certificates purpose. Default is 'server,client' for non-CA (contrary to
  only 'server' before)
- removed RC4 from default cipher suites on the server site
- refactoring of some tests using Test::More thanks to Sweet-kid and the
  2015 Pull Request Challenge
2.010 2015/01/14
- new options SSL_client_ca_file and SSL_client_ca to let the server send
  the list of acceptable CAs for the client certificate.
- t/protocol_version.t - fix in case SSLv3 is not supported in Net::SSLeay. 
  RT#101485, thanks to TEAM.
2.009 2015/01/12
- remove util/analyze.pl. This tool is now together with other SSL tools in
- added ALPN support (needs OpenSSL1.02, Net::SSLeay 1.56+) thanks to TEAM,
2.008 2014/12/16
- work around recent OCSP verification errors for revoked.grc.com (badly signed
  OCSP response, Firefox also complains about it) in test t/external/ocsp.t.
- util/analyze.pl - report more details about preferred cipher for specific TLS
2.007 2014/11/26
- make getline/readline fall back to super class if class is not sslified yet,
  i.e. behave the same as sysread, syswrite etc.
  This fixes RT#100529
2.006 2014/11/22
- Make (hopefully) non-blocking work on windows by using EWOULDBLOCK instead of
  EAGAIN. While this is the same on UNIX it is different on Windows and socket
  operations return there (WSA)EWOULDBLOCK and not EAGAIN. Enable non-blocking
  tests on Windows too.
- make PublicSuffix::_default_data thread safe
- update PublicSuffix with latest list from publicsuffix.org
2.005 2014/11/15
- next try to fix t/protocol_version.t for OpenSSL w/o SSLv3 support
2.004 2014/11/15
- only test fix: fix t/protocol_version.t to deal with OpenSSL installations
  which are compiled without SSLv3 support.
2.003 2014/11/14
- make SSLv3 available even if the SSL library disables it by default in
  SSL_CTX_new (like done in LibreSSL). Default will stay to disable SSLv3,
  so this will be only done when setting SSL_version explicitly.
- fix possible segmentation fault when trying to use an invalid certificate,
  reported by Nick Andrew.
- Use only the ICANN part of the default public suffix list and not the
  private domains. This makes existing exceptions for s3.amazonaws.com and
  googleapis.com obsolete. Thanks to Gervase Markham from mozilla.org.
2.002 2014/10/21
- fix check for (invalid) IPv4 when validating hostname against certificate. Do
  not use inet_aton any longer because it can cause DNS lookups for malformed
  IP. RT#99448, thanks to justincase[AT]yopmail[DOT]com.
- Update PublicSuffix with latest version from publicsuffix.org - lots of new
  top level domains.
- Add exception to PublicSuffix for s3.amazonaws.com - RT#99702, thanks to
2.001 2014/10/21
- Add SSL_OP_SINGLE_(DH|ECDH)_USE to default options to increase PFS security.
  Thanks to Heikki Vatiainen for suggesting.
- Update external tests with currently expected fingerprints of hosts.
- Some fixes to make it still work on 5.8.1.
2.000 2014/10/15
- consider SSL3.0 as broken because of POODLE and disable it by default.
- Skip live tests without asking if environment NO_NETWORK_TESTING is set.
  Thanks to ntyni[AT]debian[DOT]org for suggestion.
- skip tests which require fork on non-default windows setups without proper
  fork. Thanks to SHAY for https://github.com/noxxi/p5-io-socket-ssl/pull/18
1.999 2014/10/09
- make sure we don't use version 0.30 of IO::Socket::IP
- make sure that PeerHost is checked on all places where PeerAddr is
  checked, because these are synonyms and IO::Socket::IP prefers PeerHost
  while others prefer PeerAddr. Also accept PeerService additionally to
  See https://github.com/noxxi/p5-io-socket-ssl/issues/16 for details.
- add ability to use client certificates and to overwrite hostname with
1.998 2014/09/07
- make client authentication work at the server side when SNI is in by use
  having CA path and other settings in all SSL contexts instead of only the main
  one.  Based on code from lundstrom[DOT]jerry[AT]gmail[DOT]com,
1.997 2014/07/12
- thanks to return code 1 from Net::SSLeay::library_init if the library needed
  initialization and 0 if not we can now clearly distinguish if initialization
  was needed and do not need any work-arounds for perlcc by the user.
1.996 2014/07/12
- move initialization of OpenSSL-internals out of INIT again because this
  breaks if module is used with require. Since there is no right place to
  work in all circumstances just document the work-arounds needed for
  perlcc. RT#97166
1.995 2014/07/11
- RT#95452 - move initialization and creation of OpenSSL-internals into INIT
  section, so they get executed after compilation and perlcc is happy.
- refresh option for peer_certificate, so that it checks if the certificate
  changed in the mean time (on renegotiation)
- fix fingerprint checking - now applies only to topmost certificate
- IO::Socket::SSL::Utils - accept extensions within CERT_create
- documentations fixes thanks to frioux
- fix documentation bug RT#96765, thanks to Salvatore Bonaccorso.
1.994 2014/06/22
- IO::Socket::SSL can now be used as dual-use socket, e.g. start plain, upgrade
  to SSL and downgrade again all with the same object. See documentation of
  SSL_startHandshake and chapter Advanced Usage.
- try to apply SSL_ca* even if verify_mode is 0, but don't complain if this
  fails. This is needed if one wants to explicitly verify OCSP lookups even if
  verification is otherwise off, because otherwise the signature check would
  fail. This is mostly useful for testing.
- reorder documentation of attributes for new, so that the more important ones
  are at the top.
1.993 2014/06/13
- major rewrite of documentation, now in separate file
- rework error handling to distinguish between SSL errors and internal errors
  (like missing capabilities).
- fix handling of default_ca if given during the run of the program (Debian#750646)
- util/analyze-ssl.pl - fix hostname check if SNI does not work
1.992 2014/06/01
- set $! to undef before doing IO (accept, read..). On Windows a connection
  reset could cause SSL read error without setting $!, so make sure we don't
  keep the old value and maybe thus run into endless loop.
1.991 2014/05/27
- new option SSL_OCSP_TRY_STAPLE to enforce staple request even if
- work around for RT#96013 in peer_certificates
1.990 2014/05/27
- added option SSL_ocsp_staple_callback to get the stapled OCSP response
  and verify it somewhere else
- try to fix warnings on Windows again (#95967)
- work around temporary OCSP error in t/external/ocsp.t
1.989 2014/05/24
- fix #95881 (warnings on windows), thanks to TMHALL
1.988 2014/05/17
- add transparent support for DER and PKCS#12 files to specify cert and key,
  e.g. it will autodetect the format
- if SSL_cert_file is PEM and no SSL_key_file is given it will check if
  the key is in SSL_cert_file too
1.987 2014/05/17
- fix t/verify_hostname_standalone.t on systems without usable IDNA or IPv6
  #95719, thanks srchulo
- enable IPv6 support only if we have a usable inet_pton
- remove stale entries from MANIFEST (thanks seen[AT]myfairpoint[DOT]net)
1.986 2014/05/16
- allow IPv4 in common name, because browsers allow this too. But only for
  scheme www/http, not for rfc2818 (because RC2818 does not allow this).
  In default scheme IPv6 and IPv4 are allowed in CN.
  Thanks to heiko[DOT]hund[AT]sophos[DOT]com for reporting the problem.
- Fix handling of public suffix. Add exemption for *.googleapis.com
  wildcard, which should be better not allowed according to public suffix
  list but actually is used.
- Add hostname verification test based on older test of chromium. But change
  some of the test expectations because we don't want to support IP as SAN
  DNS and because we enforce a public suffix list (and thus *.co.uk should
  not be allowed)
1.985 2014/05/15
- make OCSP callback return 1 even if it was called on the server side
  because of bad setup of the socket. Otherwise we get an endless calling
  of the OCSP callback.
- consider an OCSP response which is not yet or no longer valid a soft error
  instead of an hard error
- fix skip in t/external/ocsp.t in case fingerprint does not match
- RT#95633 call EVP_PKEY_free not EVP_KEY_free in
  IO::Socket::SSL::Utils::KEY_free. Thanks to paul[AT]city-fan[DOT]org
- util/analyze.pl - with --show-chain check if chain with SNI is different
  from chain w/o SNI. 
1.984 2014/05/10
- added OCSP support:
  - needs Net::SSLeay >=1.59
  - for usage see documentation of IO::Socket::SSL (examples and anything with
    OCSP in the name)
- new tool util/analyze-ssl.pl which is intended to help in debugging of SSL
  problems and to get information about capabilities of server. Works also
  as en example of how to use various features (like OCSP, SNI..)
- fix peer_certificates (returns leaf certificate only once on client side)
- added timeout for stop_SSL (either with Timeout or with the default
  timeout for IO::Socket)
- fix IO::Socket::SSL::Utils mapping between ASN1_TIME and time_t when local
  time is not GMT. Use Net::SSLeay::ASN1_TIME_timet if available.
- fix t/external/usable_ca.t for system with junk in CA files
1.983 2014/05/03
- fix public suffix handling: ajax.googleapis.com should be ok even if googleapis.com
  is in public suffix list (e.g. check one level less)
  #95317, thanks to purification[AT]ukr[DOT]net
- usable_ca.t - update fingerprints after heartbleed attack
- usable_ca.t - make sure we have usable CA for tested hosts in CA store
1.982 2014/04/24
- fix for using subroutine as argument to set_args_filter_hack
1.981 2014/04/08
- #95432 fix ecdhe Test for openssl1.0.1d, thanks to  paul[AT]city-fan[DOT]org
- fix detection of openssl1.0.1d (detected 1.0.1e instead)
- new function can_ecdh in IO::Socket::SSL
1.980 2014/04/08
- fixed incorrect calculation of certificate fingerprint in get_fingerprint*
  and comparison in SSL_fingerprint. Thanks to
  david[DT]palmer[AT]gradwell[DOT]com for reporting.
- disable elliptic curve support for openssl 1.0.1d on 64bit because of
  openssl rt#2975
1.979 2014/04/06
- hostname checking: 
  - configuration of 'leftmost' is renamed to 'full_label', but the old
    version is kept for compatibility reasons.
  - documentation of predefined schemes fixed to match reality
1.978 2014/04/04
- RT#94424 again, fix test on older openssl version with no SNI support
1.977 2014/04/04
- fix publicsuffix for IDNA, more tests with various IDNA libs
  RT#94424. Thanks to paul[AT]city-fan[DOT]org
- reuse result of IDN lib detection from PublicSuffix.pm in SSL.pm
- add more checks to external/usable_ca.t. Now it is enough that at least
  one of the hosts verifies against the builtin CA store
- add openssl and Net::SSleay version to diagnostics in load test
1.976 2014/04/03
- added public prefix checking to verification of wildcard certificates,
  e.g. accept *.foo.com but not *.co.uk.
  See documentation of SSL_verifycn_publicsuffix and
  Thanks to noloader for pointing out the problem.
1.975 2014/04/02
- BEHAVIOR CHANGE: work around TEA misfeature on OS X builtin openssl, e.g.
  guarantee that only the explicitly given CA or the openssl default CA will
  be used. This means that certificates inside the OS X keyring will no
  longer be used, because there is no way to control the use by openssl
  (e.g. certificate pinning etc)
- make external tests run by default to make sure default CA works on all
  platforms, it skips automatically on network problems like timeouts or ssl
  interception, can also use http(s)_proxy environment variables
1.974 2014/04/02
- new function peer_certificates to get the whole certificate chain, needs
- extended IO::Socket::Utils::CERT_asHash to provide way more information,
  like issuer information, cert and pubkey digests, all extensions, CRL
  distributions points and OCSP uri
1.973 2014/03/25
- with SSL_ca certificate handles can now be used additionally to
  SSL_ca_file and SSL_ca_path
- do not complain longer if SSL_ca_file and SSL_ca_path are both given,
  instead add both as options to the CA store
- Shortcut 'issuer' to give both issuer_cert and issuer_key in CERT_create.
1.972 2014/03/23
- make sure t/external/usable_ca.t works also with older openssl without
  support for SNI. RT#94117. Thanks to paul[AT]city-fan[DOT]org
1.971 2014/03/22
- try to use SSL_hostname for hostname verification if no SSL_verifycn_name
  is given. This way hostname for SNI and verification can be specified in
  one step.
- new test program example/simulate_proxy.pl
1.970 2014/03/19
- fix rt#93987 by making sure sub default_ca does use a local $_ and not a
  version of an outer scope which might be read-only.  Thanks to gshank
1.969 2014/03/13
- fix set_defaults to match documentation regarding short names
- new function set_args_filter_hack to make it possible to override bad SSL
  settings from other code at the last moment.
- determine default_ca on module load (and not on first use in each thread)
- don't try default hostname verification if verify_mode 0
- fix hostname verification when reusing context
1.968 2014/03/13
- BEHAVIOR CHANGE: removed implicit defaults of certs/server-{cert,key}.pem
  for SSL_{cert,key}_file and ca/,certs/my-ca.pem for SSL_ca_file.
  These defaults were depreceated since 1.951 (2013/7/3).
- Usable CA verification path on Windows etc:
  Do not use Net::SSLeay::CTX_set_default_verify_paths any longer to set
  system/build dependent default verification path, because there was no
  way to retrieve these default values and check if they contained usable
  CA. Instead re-implement the same algorithm and export the results with
  public function default_ca() and make it possible to overwrite it.
  Also check for usable verification path during build.
  If no usable path are detected require Mozilla::CA at build and try to
  use it at runtime.
1.967 2014/02/06
- verify the hostname inside a certificate by default with a superset of
  common verification schemes instead of not verifying identity at all.
  For now it will only complain if name verification failed, in the future
  it will fail certificate verification, forcing you to set the expected
  SSL_verifycn_name if you want to accept the certificate.
- new option SSL_fingerprint and new methods get_fingerprint and
  get_fingerprint_bin. Together they can be used to selectively accept
  specific certificates which would otherwise fail verification, like
  self-signed, outdated or from unknown CAs.
  This makes another reason to disable verification obsolete.
- Utils:
  - default RSA key length 2048
  - digest algorithm to sign certificate in CERT_create can be given,
    defaults to SHA-256
  - CERT_create can now issue non-CA selfsigned certificate
  - CERT_create add some more useful constraints to certificate
- spelling fixes, thanks to ville[dot]skytta[at]iki[dot]fi
1.966 2014/01/21
- fixed bug introduced in 1.964 - disabling TLSv1_2 worked no longer with
  specifying !TLSv12, only !TLSv1_2 worked
- fixed leak of session objects in SessionCache, if another session 
  replaced an existing session (introduced in 1.965)
1.965 2014/01/16
- new key SSL_session_key to influence how sessions are inserted and looked
  up in the clients session cache. This makes it possible to share sessions
  over different ip:host (like required with some FTPS servers)
- t/core.t - handle case, were default loopback source is not, like
  in FreeBSD jails
1.964 2014/01/15
- Disabling TLSv1_1 did not work, because the constant was wrong. Now it gets
  the constants from calling Net::SSLeay::SSL_OP_NO_TLSv1_1 etc
- The new syntax for the protocols is TLSv1_1 instead of TLSv11. This matches
  the syntax from OpenSSL. The old syntax continues to work in SSL_version.
- New functions get_sslversion and get_sslversion_int which get the SSL version
  of the establish session as string or int.
- disable t/io-socket-inet6.t if Acme::Override::INET is installed
1.963 2014/01/13
- fix behavior of stop_SSL: for blocking sockets it now enough to call it
  once, for non-blocking it should be called again as long as EAGAIN and
- don't call blocking if start_SSL failed and downgraded socket has no 
  blocking method, thanks to tokuhirom
- documentation enhancements:
  - special section for differences to IO::Socket
  - describe problem with blocking accept on non-blocking socket
  - describe arguments to new_from_fd and make clear, that for upgrading an
    existing IO::Socket start_SSL should be used directly
1.962 2013/11/27
- work around problems with older F5 BIG-IP by offering fewer ciphers on the
  client side by default, so that the client hello stays below 255 byte
1.961 2013/11/26
- IO::Socket::SSL::Utils::CERT_create can now create CA-certificates which
  are not self-signed (by giving issuer_*)
1.960 2013/11/12
only documentation enhancements:
- clarify with text and example code, that within event loops not only
  select/poll should be used, but also pending has to be called.
- better introduction into SSL, at least mention anonymous authentication as
  something you don't want and should take care with the right cipher
- make it more clear, that user better does not change the cipher list, unless
  he really know what he is doing
1.959 2013/11/12
- bugfix test core.t windows only
1.958 2013/11/11
- cleanup: remove workaround for old IO::Socket::INET6 but instead require at
  least version 2.55 which is now 5 years old
- fix t/session.t #RT90240, thanks to  paul[AT]city-fan[DOT]org
1.957 2013/11/11
- fixed t/core.t: test uses cipher_list of HIGH, which includes anonymous
  authorization. With the DH param given by default since 1.956 old versions of
  openssl (like 0.9.8k) used cipher ADH-AES256-SHA (e.g. anonymous
  authorization) instead of AES256-SHA and thus the check for the peer
  certificate failed (because ADH does not exchanges certificates).
  Fixed by explicitly specifying HIGH:!aNULL as cipher
  RT#90221, thanks to  paul[AT]city-fan[DOT]org
- cleaned up tests: 
  - remove ssl_settings.req and 02settings.t, because all tests now create a
    simple socket at and thus global settings are no longer needed.
  - some tests did not have use strict(!), fixed it.
  - removed special handling for older Net::SSLeay versions, which are less than
    our minimum requirement
  - some syntax enhancements, removed some SSL_version and SSL_cipher_list
    options where they were not really needed
1.956 2013/11/10
lots of behavior changes for more secure defaults:
- BEHAVIOR CHANGE: make default cipher list more secure, especially
  - no longer support MD5 by default (broken)
  - no longer support anonymous authentication by default (vulnerable to man in
    the middle attacks)
  - prefer ECDHE/DHE ciphers and add necessary ECDH curve and DH keys, so that
    it uses by default forward secrecy, if underlying Net::SSLeay/openssl
    supports it
  - move RC4 at the end, e.g. 3DES is preferred (BEAST attack should hopefully
    been fixed and now RC4 is considered less safe than 3DES)
  - default SSL_honor_cipher_order to 1, e.g. when used as server it tries to
    get the best cipher even if client prefers other ciphers
  PLEASE NOTE that this might break connections with older, less secure
  implementations. In this case revert to 'ALL:!LOW:!EXP:!aNULL' or so.
- BEHAVIOR CHANGE: SSL_cipher_list now gets set on context not SSL object and
  thus gets reused if context gets reused. PLEASE NOTE that using
  SSL_cipher_list together with SSL_reuse_ctx has no longer effect on the
  ciphers of the context.
- rework hostname verification schemes
  - add rfc names as scheme (e.g. 'rfc2818',...)
  - add SIP, SNMP, syslog, netconf, GIST
  - BEHAVIOR CHANGE: fix SMTP - now accept wildcards in CN and subjectAltName
  - BEHAVIOR CHANGE: fix IMAP, POP3, ACAP, NNTP - now accept wildcards in CN
- BEHAVIOR CHANGE: anywhere wildcards like www* now match only 'www1', 'www2'..
  but not 'www'
- anywhere wildcards like x* are no longer applied to IDNA names (which start
  with 'xn--')
- fix crash of Utils::CERT_free
- support TLSv11, TLSv12 as handshake protocols
1.955 2013/10/11
- support for forward secrecy using ECDH, if the Net::SSLeay/openssl version
  supports it.
1.954 2013/9/15
- accept older versions of ExtUtils::MakeMaker and add meta information
  like link to repository only for newer versions.
1.953 2013/7/22
- fixes to IO::Socket::SSL::Utils, thanks to rurban[AT]x-ray[DOT]at,
1.952 2013/7/11
- fix t/acceptSSL-timeout.t on Win32, RT#86862
1.951 2013/7/3 
- better document builtin defaults for key,cert,CA and how they are depreceated
- use Net::SSLeay::CTX_set_default_verify_paths to use openssl's builtin
  defaults for CA unless CA path/file was given (or IO::Socket::SSL builtins
1.950 2013/7/3 
  ssl_verify_mode now defaults to verify_peer for client.
  Until now it used verify_none, but loudly complained since 1.79 about it.
  It will not complain any longer, but the connection might probably fail.
  Please don't simply disable ssl verification, but instead set SSL_ca_file
  etc so that verification succeeds!
  it will now complain if the builtin defaults of certs/my-ca.pem or ca/
  for CA and certs/{server,client}-{key,cert}.pem for cert and key are used,
  e.g. no certificates are specified explicitly.
  In the future these insecure (relative path!) defaults will be removed
  and the CA replaced with the system defaults.
v1.94 2013.06.01
- Makefile.PL reported wrong version of openssl, if Net::SSLeay was not
  installed instead of reporting missing dependency to Net::SSLeay.
v1.93 2013.05.31
- need at least OpenSSL version 0.9.8 now, since last 0.9.7 was released 6
  years ago. Remove code to work around older releases.
- changed AUTHOR in Makefile.PL from array back to string, because the
  array feature is not available in MakeMaker shipped with 5.8.9 (RT#85739)
v1.92 2013.05.30
- Intercept: use sha1-fingerprint of original cert for id into cache unless 
  otherwise given
- Fix pod error in IO::Socket::SSL::Utils RT#85733
v1.91 2013.05.30
- added IO::Socket::SSL::Utils for easier manipulation of certificates and keys
- moved SSL interception into IO::Socket::SSL::Intercept and simplified it 
  using IO::Socket::SSL::Utils
- enhance meta information in Makefile.PL
v1.90 2013.05.27
- RT#85290, support more digest, especially SHA-2.
  Thanks to ujvari[AT]microsec[DOT]hu
- added support for easy SSL interception (man in the middle) based
  on ideas found in mojo-mitm proxy (which was written by Karel Miko)
- make 1.46 the minimal required version for Net::SSLeay, because it 
  introduced lots of useful functions.
v1.89 2013.05.14
- if IO::Socket::IP is used it should be at least version 0.20, otherwise
  we get problems with HTTP::Daemon::SSL and maybe others (RT#81932)
- Spelling corrections, thanks to dsteinbrunner
v1.88 2013.05.02
- consider a value of '' the same as undef for SSL_ca_(path|file), SSL_key*
  and SSL_cert* - some apps like Net::LDAP use it that way.
  Thanks to alexander[AT]kuehn[AT]nagilum[DOT]de for reporting the problem.
v1.87 2013.04.24
- RT#84829 - complain if given SSL_(key|cert|ca)_(file|path) do not exist or
  if they are not readable. Thanks to perl[AT]minty[DOT]org 
- fix use of SSL_key|SSL_file objects instead of files, broken with 1.83
v1.86 2013.04.17
- RT#84686 - don't complain about SSL_verify_mode is SSL_reuse_ctx,
  thanks to CLEACH
v1.85 2013.04.14
- probe for available modules with local __DIE__ and __WARN__handlers.
  fixes RT#84574, thanks to FRAZER
- fix warning, when IO::Socket::IP is installed and inet6 support gets explicitly
  requested. RT#84619, thanks to Prashant[DOT]Tekriwal[AT]netapp[DOT]com
v1.84 2013.02.15
- disabled client side SNI for openssl version < 1.0.0 because of RT#83289
- added functions can_client_sni, can_server_sni, can_npn to check availability
  of SNI and NPN features. Added more documentation for SNI and NPN.
v1.83_1 2013.02.14
- separated documentation of non-blocking I/O from error handling
- changed and documented behavior of readline to return the read
  data on EAGAIN/EWOULDBLOCK in case of non-blocking socket.
  See https://github.com/noxxi/p5-io-socket-ssl/issues/1, thanks to
v1.83 2013.02.03
- Server Name Indication (SNI) support on the server side, inspired by
  patch provided by karel[DOT]miko[AT]gmail[DOT]com.
- reworked part of the documentation, like providing better examples.
v1.82 2013.01.28
- sub error sets $SSL_ERROR etc only if there really is an error, 
  otherwise it will keep the latest error. This causes 
  IO::Socket::SSL->new.. to report the correct problem, even if
  the problem is deeper in the code (like in connect)
- correct spelling, rt#8270. Thanks to ETHER
v1.81 2012.12.06
- deprecated set_ctx_defaults, new name ist set_defaults (but old name
  still available)
- changed handling of default path for SSL_(ca|cert|key)* keys: either
  if one of these keys is user defined don't add defaults for the
  others, e.g.  don't mix user settings and defaults
- cleaner handling of module defaults vs. global settings vs. socket
  specific settings. Global and socket specific settings are both
  provided by the user, while module defaults not.
- make IO::Socket::INET6 and IO::Socket::IP specific tests run both,
  even if both modules are installed by faking a failed load of the
  other module.
v1.80 2012.11.30
- removed some warnings in test (missing SSL_verify_mode => 0) which
  caused tests to hang on Windows.
v1.79 2012.11.25
- prepare transition to a more secure default for SSL_verify_mode.
  The use of the current default SSL_VERIFY_NONE will cause a big warning
  for clients, unless SSL_verify_mode was explicitly set inside the 
  application to this insecure value.
  In the near future the default will be SSL_VERIFY_PEER, and thus 
  causing verification failures in unchanged applications.
v1.78 2012.11.25
- use getnameinfo instead of unpack_sockaddr_in6 to get PeerAddr and 
  PeerPort from sockaddr in _update_peer, because this provides scope
  too. Thanks to bluhm[AT]genua[DOT]de.
- work around systems which don't defined AF_INET6
  Thanks to GAAS for reporting
v1.77 2012.10.05
- update_peer for IPv6 also, applied fix to 
  https://rt.cpan.org/Ticket/Display.html?id=79916 by
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
v1.74_2 2012.06.07
- fixed documentation errors, reported by MARSCHAP
v1.74_1 2012.06.07
- add support to IO::Socket::IP which support inet6 and inet4 by
  integrating patch from PEVANS
  for https://rt.cpan.org/Ticket/Display.html?id=75218
v1.74 2012.05.13
- accept a version of SSLv2/3 as SSLv23, because older documentation
  could be interpreted like this
v1.73 2012.05.11
- make test t/dhe.t hopefully work for more version of openssl
  Thanks to paul[AT]city-fan[DOT]org for providing bug reports and
  testing environment
v1.72 2012.05.10
  Thanks to dcostas[AT]gmail[DOT]com for problem report
v1.71 2012.05.09
- 1.70 done right. Also don't disable SSLv2 ciphers, SSLv2 support is better
  disabled by the default SSL_version of 'SSLv23:!SSLv2'
v1.70 2012.05.08
- make it possible to disable protocols using SSL_version, make SSL_version
  default to 'SSLv23:!SSLv2'
v1.69 2012.05.08
- re-added workaround in t/dhe.t
v1.68 2012.05.07
- remove SSLv2 from default cipher list, which makes failed tests after last
  change work again, fix behavior for empty cipher list (use default)
v1.67 2012.05.07
- https://rt.cpan.org/Ticket/Display.html?id=76929
  thanks to d[DOT]thomas[AT]its[DOT]uq[DOT]edu[DOT]au for reporting
  - if no explicit cipher list is given it will now default to ALL:!LOW instead 
    of the openssl default, which usually includes weak ciphers like DES.
  - new config key SSL_honor_cipher_order and documented how to use it to fight
    BEAST attack. 
v1.66 2012.04.16
- make it thread safer, thanks to bug report from vega[DOT]james[AT]gmail
  [DOT]com, https://rt.cpan.org/Ticket/Display.html?id=76538
v1.65 2012.04.16
- added NPN (Next Protocol Negotiation) support based on patch from kmx
v1.64 2012.04.06
- clarify some behavior regarding hostname verification.
  Thanks to DOHERTY for reporting.
v1.63 2012.04.06
- applied patch of DOUGDUDE to ignore die from within eval to make tests 
  more stable on Win32, https://rt.cpan.org/Ticket/Display.html?id=76147
v1.62 2012.03.28
- small fix to last version
v1.61 2012.03.27
- call CTX_set_session_id_context so that servers session caching works with
  client certificates too. 
v1.60 2012.03.20
- don't make blocking readline if socket was set nonblocking, but return as
  soon no more data are available
- fix BUG section about threading so that it shows package as thread safe
  as long as Net::SSLeay >= 1.43 is used
v1.59 2012.03.08
- if SSLv2 is not supported by Net::SSLeay set SSL_ERROR with useful 
  message when attempting to use it. 
- modify constant declarations so that 5.6.1 should work again
v1.58 2012.02.26
- fix t/dhe.t again to enable the workaround only for newer openssl
  versions, because this would cause failures on older versions
v1.57 2012.02.26
- fix t/dhe.t for openssl 1.0.1 beta by forcing tlsv1, so that it does
  not complain about the too small rsa key which it should not use anyway.
  Thanks to paul[AT]city-fan[DOT]org  for reporting.
v1.56 2012.02.22
- add automatic or explicit (via SSL_hostname) SNI support, needed for
  multiple SSL hostnames with same IP. Currently only supported for the 
v1.55 2012.02.20
- work around IO::Sockets work around for systems returning EISCONN etc
  on connect retry for non-blocking sockets by clearing $! if SUPER::connect
  returned true.
  Thanks for Manoj Kumar for reporting.
v1.54 2012.01.11
- return 0 instead of undef in SSL_verify_callback to fix uninitialized
  warnings.  Thanks to d[DOT]thomas[AT]its[DOT]uq[DOT]edu[DOT]au for 
  reporting the bug and MIKEM for the fix.
v1.53 2011.12.11
- kill child in t/memleak_bad_handshake.t if test fails
  Thanks to CLEACH ofr reporting
v1.52 2011.12.07
- fix syntax error in t/memleak_bad_handshake.t
  thanks to cazzaniga[DOT]sandro[AT]gmail[DOT]com for reporting
v1.51 2011.12.06
- disable t/memleak_bad_handshake.t on AIX, because it might hang
v1.50 2011.12.06
  Thanks to HMBRAND for reporting and Rainer Tammer tammer[AT]tammer[DOT]net for
  providing access to AIX system
v1.49 2011.10.28 
- another regression for readline fix, this time it failed to return lines
  at eof which don't end with newline. Extended t/readline.t to catch this
  case and the fix for 1.48
  Thanks to christoph[DOT]mallon[AT]gmx[DOT]de for reporting
v1.48 2011.10.26
- bugfix for readline fix in 1.45. If the pending data where false
  (like '0') it failed to read rest of line.
  Thanks to Victor Popov for reporting
v1.47 2011.10.21
- fix for 1.46 - check for mswin32 needs to be /i. Thanks to
  Alexandr Ciornii for reporting
v1.46 2011.10.18
- disable test t/signal-readline.t on windows, because signals are
  not relevant for this platform and test does not work.
v1.45 2011.10.12
- fix readline to continue when getting interrupt waiting for more
  data. Thanks to kgc[AT]corp[DOT]sonic[DOT]net for reporting problem
v1.44 2011.05.27
- fix invalid call to inet_pton in verify_hostname_of_cert when 
  identity should be verified as ipv6 address, because it contains
v1.43_1 2011.05.12
- try to make t/nonblock.t more stable, especially on Mac OS X
v1.43 2011.05.11
- fix t/nonblock.t
- stability improvements t/inet6.t
v1.42 2011.05.10
- add SSL_create_ctx_callback to have a way to adjust context on
  creation. https://rt.cpan.org/Ticket/Display.html?id=67799
- describe problem of fake memory leak because of big session cache
  and how to fix it, see https://rt.cpan.org/Ticket/Display.html?id=68073
v1.41 2011.05.09
- fix issue in stop_SSL where it did not issue a shutdown of the 
  SSL connection if it first received the shutdown from the other
  side. Thanks to fencingleo[AT]gmail[DOT]com for reporting
- try to make t/nonblock.t more reliable, at least report the real
  cause of ssl connection errors
v1.40 2011.05.02
- integrated patch from GAAS to get IDN support from URI.
v1.39_1 2011.05.02
- fix in exampel/async_https_server.
  Thanks to DetlefPilzecker[AT]web[DOT]de for reporting
v1.39 2011.03.03
- fixed documentation of http verification: wildcards in cn is allowed
v1.38_1 2011.01.24
- close should undef _SSL_fileno, because the fileno is no longer
  valid (SSL connection and socket are closed)
v1.38 2011.01.18
- fixed wildcards_in_cn setting for http (wrongly set in 1.34 to 1
  instead of anywhere). Thanks to dagolden[AT]cpan[DOT]org for
v1.37 2010.12.09
- don't complain about invalid certificate locations if user explicitly 
  set SSL_ca_path and SSL_ca_file to undef. Assume that user knows what
  he is doing and will work around the problems by itself.
v1.36 2010.12.08
- update documentation for SSL_verify_callback based on 
v1.35 2010.12.06
- if verify_mode is not VERIFY_NONE and the ca_file/ca_path cannot be
  verified as valid it will no longer fall back to VERIFY_NONE but throw
  an error. Thanks to Salvatore Bonaccorso and Daniel Kahn Gillmor for
  pointing out the problem, see also 
v1.34 2010.11.01
- scheme http for certificate verification changed to 
  wildcards_in_cn=1, because according to rfc2818 this is valid and
  also seen in the wild
- if upgrading socket from inet to ssl fails due to handshake problems
  the socket gets downgraded, but is still open.
  See https://rt.cpan.org/Ticket/Display.html?id=61466
- deprecate kill_socket, just use close()
v1.33 2010.03.17
- attempt to make t/memleak_bad_handshake.t more stable, it fails 
  for unknown reason on various systems
- fix hostname checking: an IP should only be checked against 
  subjectAltName GEN_IPADD, never against GEN_DNS or CN.
  Thanks to rusch[AT]genua[DOT]de for bug report
v1.32 2010.02.22
- Makefile.PL: die if Scalar::Util has no dualvar support instead of
  only complaining. Thanks to w[DOT]phillip[DOT]moore[AT]gmail[DOT]com
  for reporting.
v1.31 2009.09.25
- add and export constants for SSL_VERIFY_* 
- set SSL_use_cert if cert is given and not SSL_server
- support alternative CRL file with SSL_crl_file thanks to patch of
v1.30_3 2009.09.03
- make t/memleak_bad_handshake.t more stable (increase listen queue,
  ignore errors on connect, don't run on windows..)
v1.30_2 2009.09.01
- t/memleak_bad_handshake.t don't write errors with ps to stderr,
  -o vsize argument is not supported on all platforms, just skip
  test then
v1.30_1 2009.08.31
- make sure that idn_to_ascii gets no \0 bytes from identity, because
  it simply cuts the string their (using C semantics). Not really a
  security problem because IDN like identity is provided by user in
  hostname, not by certificate.

v1.30 2009.08.19
- fix test t/memleak_bad_handshake.t
v1.29 2009.08.19
- fixed thanks for version 1.28
v1.28 2009.08.19
- fix memleak when SSL handshake failed.
  Thanks richardhundtu[AT]gmail[DOT]com

v1.27 2009.07.24
- changed possible local/utf-8 depended \w in some regex against more
  explicit [a-zA-Z0-9_]. Fixed one regex, where it assumed, that service
  names can't have '-' inside
- fixed bug https://rt.cpan.org/Ticket/Display.html?id=48131
  where eli[AT]dvns[DOT]com reported warnings when perl -w was used.
  While there made it more aware of errors in Net::ssl_write_all (return
  undef not 0 in generic_write)
v1.26 2009.07.03
  fix Bug in verify_hostname_of_cert where it matched only the prefix for 
  the hostname when no wildcard was given, e.g. www.example.org matched
  against a certificate with name www.exam in it
  Thanks to MLEHMANN for reporting

v1.25 2009.07.02
- t/nonblock.t: increase number of bytes written to fix bug with OS X 10.5

v1.24 2009.04.01
- add verify hostname scheme ftp, same as http
- renew test certificates again (root CA expired, now valid for 10 years)

v1.23 2009.02.23
- if neither SSL_ca_file nor SSL_ca_path are known (e.g not given and the
  default values have no existing file|path) disable checking of
  certificates, but carp about the problem
- new test certificates, the old ones expired and caused tests to fail

v1.22 2009.01.24
- Net::SSLeay stores verify callbacks inside hash and never clears them, so
  set verify callback to NULL in destroy of context

v1.21 2009.01.22
- auto verification of name in certificate created circular reference between
  SSL and CTX object with the verify_callback, which caused the objects to be
  destroyed only at program end. Fix it be no longer access $self from inside
  the callback.
  Thanks to odenbach[AT]uni-paderborn[DOT]de for reporting

v1.20 2009.01.15
- only changes on test suite to make it ready for win32
  (tested with strawberry perl 5.8.8)

v1.19 2008.12.31
- fix verifycn_name autodetection from PeerAddr/PeerHost

v1.18 2008.11.17
- fixed typo in argument: wildcars_in_cn -> wildcards_in_cn
  thanks to ludwig[DOT]nussel[AT]suse[DOT]de for reporting

v1.17 2008.10.13
- no code changes, publish v.16_3 as v.17 because it looks better 
  than v.16
- document win32 behavior regarding non-blocking and timeouts

v1.16_3   2008.09.25
- fix t/nonblock.t with workaround for problems with 
  IO::Socket::INET on some systems (Mac,5.6.2) where it cannot do 
  nonblocking connect and leaves socket blocked.
- make some tests less verbose by fixing diag in t/testlib.t 
  (send output to STDOUT not STDERR and prefix with '#')

v1.16_2   2008.09.24
- work around Bug in IO::Socket::INET6 on BSD systems
  by setting Domain based on PeerAddr
  Thanks to srezic for report and support
- remove tests of recv/send from t/core.t. Might badly interact
  with SSL handshake and cause crashes as seen on OS X 10.4

v1.16_1   2008.09.19
- better support for IPv6:
  - IPv6 is enabled by default if IO::Socket::INET6 is available
  - t/inet6.t for basic tests

v1.16    2008.09.19
- change code for SSL_check_crl to use X509_STORE_set_flags instead of
  X509_STORE_CTX_set_flags based on bug report from 
  <tjtoocool[AT]phreaker[DOT]net >
- change opened() to report -1 if the IO::Handle is open, but the
  SSL connection failed, needed with HTTP::Daemon::SSL which will send
  an error message over the unencrypted socket

- change internal behavior when SSL handshake failed (like when verify
  callback returned an error) in the hope to fix spurious errors in

- added support for verification of hostname from certificate
  including subjectAltNames, support for IDN etc based on patch and
  input from christopher[AT]odenbachs[DOT]de and 
  It is also possible to get more information from peer_certificate
  based on this patch. See documentation for peer_certificate and
- automatic verification of hostnames with SSL_verifycn_scheme and
- global setting of default context options like SSL_verifycn_scheme,
  SSL_verify_mode with set_ctx_defaults
- fix import of inet4,inet6 which got broken within 1.13_X.
  Thanks to <at[AT]altlinux[DOT]ru> for bugreport and patch
- clarified and enhanced debugging support based on bugreport
- put information into README regarding the supported and recommended
  version of Net::SSLeay

- removed CLONE_SKIP which was added in 1.03 because this breaks
  windows forking. Handled threads/windows forking better by making
  sure that CTX from Net::SSLeay gets not freed multiple times from 
  different threads after cloning/forking
- removed setting LocalPort to 0 in tests, instead leave it undef
  if a random port should be allocated. This should fix build problems 
  with 5.6.1. Thanks to <andrew[DOT]benham[AT]thus[DOT]net>

- treat timeouts of 0 for accept_SSL and connect_SSL like no timeout,
  like IO::Socket does.

- fixed errors in accept_SSL which would work when called from start_SSL
  but not from accept

- start_SSL, accept_SSL and connect_SSL have argument for Timeout
  so that the SSL handshake will not block forever. Only used if the
  socket is blocking. If not set the Timeout value from the underlying
  IO::Socket is used

- new method stop_SSL as opposite of start_SSL based on a idea
  of Bron Gondwana <brong[AT]fastmail[DOT]fm>
  To support this method the SSL_shutdown handling had to be
  fixed, e.g. in close a proper unidirectional shutdown
  should be done while in stop_SSL a bidirectional shutdown
- try to make it clearer that thread support is buggy

- make sure that Scalar::Util has support for dualvar
  (Makefile.PL,SSL.pm) because the perl-only version has
  has no dualvar

- fix t/nonblock.t on systems which have by default a larger
  socket buffer. Set SO_SNDBUF explicitly with setsockopt
  to force smaller writes on the socket

- instead of setting undef args to '' in configure_SSL drop
  them. This makes Net::SMTP::SSL working again because it
  does not give LocalPort of '' to IO::Socket::INET any more

- make session cache working even if the IO::Socket::SSL object
  was not created with IO::Socket::SSL->new but with
  IO::Socket::SSL->start_SSL on an established socket

- added way to create SSL object with predefined session
  cache, thus making it possible to share the cache between
  objects even if the rest of the context is not shared
  key SSL_session_cache
  Note that the arguments of IO::Socket::SSL::SessionCache::new
  changed (but you should never have used this class directly
  because it's internal to IO::Socket::SSL)

- add CLONE_SKIP as proposed by 
  Jarrod Johnson jbjohnso at us dot ibm dot com

- added some info to BUGS and to BUGS section of pod
- added TELL and BINMODE to IO::Socket::SSL::SSL_HANDLE, even
  if they do nothing useful.
- all tests allocate now the ports dynamically, so there should
  be no longer a conflict with open ports on the system where
  the tests run

- work around Bug in Net::HTTPS where it defines sub blocking
  as {}, e.g. force scalar context when calling sub blocking
  (in IO::Socket::SSL::write)
  see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=383106

- fix deprecated and practically undocumented function 
  get_peer_certificate so that LWP Net::HTTPS works again
- set arg 'Blocking' while calling SUPER::configure only
  if it was set by the caller to work around Problem in LWP

- If SSL_cipher_list is not given it uses the openssl
  default instead of setting it to 'ALL:!LOW:!EXP' like
  before. The old value included ADH and this might be
  a bad idea, see BUGS why.

- declare socket as opened before calling fatal_ssl_error
  because the SSL_error_trap set up from HTTP::Daemon
  needs this
- accept_SSL sets errors on $socket (the accepted socket)
  not $self (the listening socket if called from accept)
  so it can be queried from SSL_error_trap
- note in BUGS section that IO::Socket::SSL is not thread-safe

- fix readline (e.g. getline,getlines,<>) so that it behaves
  regarding $/ like written in the $/ documentation.

- removed links and comments to unofficial release of
  Net::SSLeay, because there is a newer version already

- add support for Diffie Hellman Key Exchange.
  See parameter SSL_dh_file and SSL_dh.

- hide DEBUG statements and remove test to load Debug.pm
  because packages like SpamAssassin cannot cope with it
  (at least the OpenBSD port)

- added SSL_cert and SSL_key parameter which do not take
  a file name like SSL_cert_file and SSL_key_file but
  an internal X509* resp. EVP_PKEY* value. Useful for
  dynamically created certificates and keys.
- added test for sysread/syswrite behavior (which was changed
  in v0.991)

- _set_rw_error does $!||=EAGAIN only if error is one of 
  SSL_WANT_READ|SSL_WANT_WRITE (patch from Mike Smith 
  <mike at mailchannels dot com>)
- Fix Makefile.PL to allow detection of failures in PREREQ_PM
  (http://rt.cpan.org/Public/Bug/Display.html?id=20563, patch 
  by alexchorny at gmail dot com)

- sysread and syswrite ar no longer the same as read and write,
  but can return already if only parts of the data are read
  or written (which is the usual semantic for sysread and syswrite)
  This should fix problems with HTTP::Daemon::SSL

- just upgrade Version number because I've screwed up upload
  of v0.98 to cpan

- Maintainer changed to <Steffen_Ullrich at genua dot de>
- Better support for nonblocking sockets:
  . exports $SSL_ERROR which contains the latest error from
    the openssl library. Exports constants SSL_WANT_READ and
    SSL_WANT_WRITE es special errors which will be set if
    openssl wants to write or read during nonblocking connects,
    accepts, reads or writes.
  . accept,accept_SSL,connect and connect_SSL don't block
    anymore if the socket is nonblocking.
    Instead $! will be set from the underlying IO::Socket::INET
    connect or accept if it failed there (usually EAGAIN or 
    EINPROGRESS) or if the underlying openssl needs to read or
    write $! will be set to EAGAIN and $SSL_ERROR will be set
  . syswrite returns undef and sets $!,$SSL_ERROR if it fails 
    to write instead of returning 0.
- Bugfixes (http://rt.cpan.org/Public/Bug/Display.html?id=Bugid)
  . Bug 18439: fileno 0 should be valid
  . Bug 15001: sysread interprets buffer "0" as ""
- peer_certificate returns X509 struct string if no field
  for extraction was specified 
- get_peer_certificate returns the certificate instead of the
  IO::Socket::SSL object

- Writes now correctly return errors.  (Problem noted by
  Dominique Quatravaux <dom at idealx.com>).
- CA paths now work without passing an empty SSL_ca_file
  argument.  (Problem found by Phil Pennock, <phil.pennock
  at globnix.org>).
- IO::Socket::SSL now automatically passes Proto => tcp (if
  not already specified) to IO::Socket::INET to work around
  /etc/services files with udp entries listed first.  (Fix
  suggested by Phil Pennock).
- $socket->accept() now returns the peer address in array
  context for better conformance with IO::Socket::INET.
  However, if you were doing "map { $_->accept } (@sockets)",
  or similar tricks, you will need to use "scalar" to get the
  old behavior back.  (Problem noted by Nils Sowen, <n.sowen
  at kon.de>).
- IO::Socket::SSL should now properly block on reads larger
  than the buffer size of Net::SSLeay.  (Problem found by Eric
  Jergensen, <eric at dvns.com>).
- IO::Socket::SSL should now send CA Certs (if necessary)
  along with certificates.  (Problem found by <roy at
- Timeouts should now work, but be aware that if multiple
  reads/writes are necessary to complete a connection, then
  each one may have a separate timeout.  (Request from
  Dominique Quatravaux <dom at idealx.com>).
- In certain cases, start_SSL() would misplace a socket's
  fileno, causing problems with starting SSL.  This should now
  be fixed.  (Problem found by <russ at zerotech.net>).
- IO::Socket::SSL now requires a minimum of Net::SSLeay 1.21.

--- Old Versions --------------------------------------------------

v0.96  2004.4.30
- Makefile's error messages now correct if output is
  redirected (patch from Ilya Zakharevich <ilya at
- Non-blocking connects/accepts now work (Problem found by
  Uri Guttman <uri at stemsystems.com>).
- new_from_fd() now works.
- getline() and <> in scalar context now return undef
  instead of '' if the read failed.  (Problem found by
  Christian Gilmore <cag at us.ibm.com>).
- Broken pipe signals are now ignored during socket close
  to prevent a SSL shutdown message from killing the parent
  program.  (Problem found by Christian Gilmore).
- Tests should proceed much more quickly, and a semi-race was
  fixed, meaning that on slow machines the tests should be
  more reliable.
- Check for Scalar::Util and Weakref now uses default
  $SIG{__DIE__} instead of a potentially user-altered one
  (suggestion from Olaf Schneider <Olaf.Schneider at
  iwr.fzk.de>).  This only applies to Perl 5.6.0 & above.
- Session caching support (patch from Marko Asplund
  <marko.asplund at kronodoc.fi>).
- set_default_context() added to alter the behavior of
  modules that use IO::Socket::SSL from the main program.
- get_ssl_object() renamed to _get_ssl_object() to reflect
  the fact that it's only supposed to be used internally
  (not that you should have cared, of course).
- Added patch for Net::SSLeay to take advantage of
  client-side session caching.

v0.95  2003.8.25
- Changed PeerAddr in example/ssl_client.pl back to localhost.
- Update of examples to automatically switch to the proper
  directory if they cannot find the necessary SSL certificates.
- Minor documentation update with more INET6 info.
- Corrected some error messages for IO::Socket::INET6.
- Better opened() behavior when sockets close unexpectedly.
- Added note about random number generators for Solaris users
  (Problem found by Christian Gilmore <cag at us.ibm.com>).
- Added support for WeakRef and Scalar::Util to allow
  IO::Socket::SSL objects to auto-destroy themselves when 
  they go out of scope.
- Added croak()ing for unimplemented send() and recv() methods
  so they are not accidentally used to transmit unencrypted
  data.  The Perl builtin functions cannot be reliably trapped
  and are still dangerous, a fact that the POD now reflects
  (Problem noted by Michal Ludvig <michal at logix.cx>).

v0.94  2003.6.26
- Changed accept() to use inherited accept() instead of
  IO::Socket::accept, so that IPv6 inheritance is possible.
- Added options to import() so that a user could specify
  IPv6 or IPv4 mode of operation.
- Documentation fixes, esp. e-mail address.

v0.93  2003.6.24
- Fixed error-checking slip in connect_SSL() (Problem found by
  Uri Guttman <uri at stemsystems.com>).
- All functions now return the empty list () on errors.
- Added note about the above change to appease Graham Barr
  <gbarr at pobox.com>.
- Fixed Net::SSLeay giving warnings when arguments are undef;
  in all cases, undef arguments may be set to '' without any
  change in behavior except for removing the warnings.
  (Problem found by Dominique Quatravaux <dom at idealx.com>)
- If accept() or connect() fails in SSL negotiation, the user
  now has the option to print something to the failed socket
  before it is closed.  (error_trap option in new())
- Added support for CRLs (SSL_check_crl option in new()) for
  versions of OpenSSL >= 0.9.7b (Original patch from
  Brian Lindauer <jbl at sysd.com>)
- Finally added decent support for certificate callbacks.
  (SSL_verify_callback option in new(), suggestion from
  Dariush Pietrzak <eyck at ghost.anime.pl>).
- accept()/connect()/socket_to_SSL() now fail immediately if
  the socket in question does not have a fileno.
- Added the kill_socket() method to guarantee that a socket dies.
- Fixed extra warning when printing errors in debug mode.
- Deprecated socket_to_SSL() in favor of the class method
  start_SSL() (Class method suggestion from Graham Barr
  <gbarr at pobox.com>).
- Added the class method start_SSL() to allow for cases when
  the desired class of the socket is not IO::Socket::SSL
  (Request from Dariush Pietrzak <eyck at ghost.anime.pl>)
- Changed socket_to_SSL to rebless socket to original class
  if SSL negotiation failed (Request from Graham Barr
  <gbarr at pobox.com>)
- Removed the daemon.pl example, as it did not work with the
  standard distribution of HTTP::Daemon (use HTTP::Daemon::SSL

v0.92   2002.10.22
- Changed the fileno() function to support returning the fileno
  of server sockets.  (Problem found by Roland Giersig
  <RGiersig at cpan.org>).
- Fixed SSL_version incorrectly defaulting to SSLv2 (patch from
  Roland Alder <roland.alder at celeris.ch>).

v0.91   2002.08.31
- Added support for SSL_peek and SSL_pending (peek() and
  pending()).  Updated documentation, tests, etc. to reflect

v0.901  2002.08.19
- Fixed the warning that happens when sockets are not explicitly
  closed() before the program terminates.

v0.90   2002.08.13
- This version is a complete rewrite of IO::Socket::SSL.  It now
  has about half the lines of code, twice the amount of documentation,
  and a slightly more polished interface.
- IO::Socket::SSL now works properly with mod_perl and taint mode.
- Major documentation update.
- Update of the BUGS file to reflect changes made in the rewrite.
- Update of the test suite for Perl v5.8.0 (or, more precisely,
  for Scalar::Util).
- Update of the test suite for Perl v5.00503 (or, more precisely,
  for the lack of several nice features added in v5.6.0) (Marko
  Asplund <aspa at kronodoc.fi>).
- New test suite that does not need the Internet to function.
- Update of all the files in example/ to use more current features
  of IO::Socket::SSL.
- Removal of SSL_SSL and X509_Certificate classes.
- There have been a few name changes (like socketToSSL ->
  socket_to_SSL) for better consistency.
- The functionality of get_peer_certificate() and friends is deprecated.
- The functionality of want_write() and want_read() is deprecated.
- The functionality of context_init() is deprecated for normal use.
- Support for all SSL context options in the new() call.
- SSL contexts are no longer global.  The SSL_reuse_ctx option
  is provided for those who want to re-use a context.
- The default verify mode is now VERIFY_NONE.
- IO::Socket::SSL::DEBUG is now linked to Net::SSLeay::trace to
  provide different levels of debugging information.
- There is a uniform interface for error reporting, so on error
  all functions will return undef and the error will be available
  by calling errstr().
- The dump_peer_certificate() and peer_certificate() functions
  have been added.
- sysread() will now behave correctly if the offset argument is
  greater than the length of the read buffer.  It also will truncate
  the read buffer properly, according to the Perl documentation for
- getline(), getlines(), and getc() have been added.
- syswrite() now uses references to avoid copying large	amounts of data.
- readline() uses ssl_read_all in array context for improved speed.
- close() now uses SSL_shutdown() to properly close an SSL connection,
  unless you tell it not to.
- If you have Net::SSLeay version 1.18 or greater, X509 certificates
  will be properly freed.
- All other known bugs have been fixed.

v0.81a (Not publicly released)
- Added support for SSL_passwd_cb.
- Added accept() server socket support to socketToSSL().

v0.81   2002.04.10
- calling context_init twice destroyed global context. fix from
  Jason Heiss <jheiss at ofb.net>.
- file handle tying interface implementation moved to a separate
  class to prevent problems resulting from self-tying filehandles.
  Harmon S. Nine <hnine at netarx.com>.
- docs/debugging.txt file added
- require Net::SSLeay v1.08
- preliminary support for non-blocking read/write
- socketToSSL() now respects context's SSL verify setting
  reported by Uri Guttman <uri at stemsystems.com>.

v0.80	2001.08.19
- fixed startTLS support (socketToSSL) (Graham Barr <gbarr at pobox.com>)
- make accept() set fileno attribute on newly created IO::Socket::SSL
  object (Martin Oldfield <m at mail.tc>).
- certificate updates.
- use SSL_CTX_use_PrivateKey_file in SSL_Context::new.

v0.79	2001.06.04
- angle bracket readline operator support
  (David Darville <david at dark.x.dtu.dk>).
- eliminate warnings in choosing SSL protocol version.
- implement our own opened method and make length parameter optional
  in syswrite (Robert Bihlmeyer <robbe at orcus.priv.at>).

v0.78	2001.04.24
- test script targets changed, certificate setup fixed
- support for TLS in SSL_version. SSL_version parameter values
  changed from integer to string. NB: this is an incompatible change.
  all SSL_version parameter values have to be changed. valid values
  include: 'sslv2', 'sslv3', 'sslv23'. Stephen C. Koehler
  <koehler at securecomputing.com>.
- enable selecting SSL version for connections. patch from
  Takanori Ugai <ugai at jp.fujitsu.com>.
- allow setting SSL_ca_file to ''. this is needed for being
  able to use SSL_ca_path (Robert Bihlmeyer <robbe at orcus.priv.at>).
- include the Apache CA bundle file in the distribution (my-ca.pem).
- BUGS file added.

v0.77	2001.01.15
- don't setup SSL CA verification unless cert verification is
  actually used for the connections.
- default SSL protocol version selection in SSL.pm.

v0.76	2000.11.17
- patch from Kwok Chern Yue <chernyue at post1.com> for
  making IO::Socket::SSL work with HTTP::Daemon.

v0.75	2000.07.26
- IO::Socket::SSL should now work with perl v5.6.0
- demo/*.pl and t/*.t now turn module debugging on if
  DEBUG command line argument is given
- default certificates changed

v0.74	2000.07.05
- Changes file added
- bugfix in IO::Socket::SSL::sysread() (zliu2 at acsu.buffalo.edu)
- libwww-perl and IO::Socket::SSL UML models added in docs
- URL changes in test scripts
- preliminary support for startTLS in IO::Socket::SSL::socketToSSL()
- miscellaneous patches for Net::SSLeay added in diffs