OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
Re: aliases problem

From: Victor Duchovni (Victor.DuchovniMorganStanley.com)
Date: Tue Feb 14 2006 - 08:55:59 CST


On Tue, Feb 14, 2006 at 10:42:48AM -0400, Peter Marshall wrote:

> postconf -d | grep version
> mail_version = 2.2.1
>

Download and install 2.2.8 (or if you can wait a little bit, 2.2.9).
There are no changes to the RELEASE_NOTES between 2.2.1 and 2.2.9-RC1.
No major or incompatible changes happen in the official release.

The following issues are solved between 2.2.1 and 2.2.9-RC1

    20050317
    
           Cleanup: change wording of error message when an IPv6 address
           is mistaken for maptype:mapname. File: util/dict_open.c.
    
    20050321
    
           Robustness: don't look for SMTP status code when there was
           none. File: smtp/smtp_chat.c, lmtp/lmtp_chat.c.
    
           Portability: missing netinet/in.h include, so that ntohs()
           was not defined on HP-UX. File: smtp/smtp_proto.c.
    
    20050327
    
           Bugfix: the SMTP and LMTP clients did not ask the queue
           manager to reduce destination concurrency when "lost
           connection" or "connection timed out" happened AFTER Postfix
           received the server greeting. Files: smtp/smtp_trouble.c,
           lmtp/lmtp-trouble.c.
    
    20050328
    
           Cleanup: the REPLACE action is no longer implemented as
           PREPEND+IGNORE. The result remains in the input stream, and
           is subject to address rewriting and other processing where
           applicable. File: cleanup/cleanup_message.c.
    
    20050402
    
           Cleanup: updated error messages about MIME processing errors
           in the SMTP client. These errors are no longer specific to
           8bit->7bit conversion; they can also happen with generic
           address mapping. File: smtp/smtp_proto.c.
    
    20050407
    
           %^!#& Thanks to inadequate SASL documentation the client
           could negotiate a security layer where none was desired.
           Better documentation has become available since Postfix
           SASL support was implemented, and now Postfix needs to be
           fixed. Files: */*_sasl_glue.c.
    
    20050409
    
           Safety: the CDB map now logs a warning when the source file
           is newer than the indexed file, just like the Berkeley DB
           and DBM maps. Michael Tokarev. File: util/dict_cdb.c.
    
    20050412
    
           Bugfix: while updating the cleanup_flush() infrastructure
           in the 2.3 development release, eliminated a portability
           problem that was introduced when "REJECT text" support was
           added. File: cleanup/cleanup.c.
    
    20050413
    
           Portability: don't mix socket message send/receive calls
           with socket stream read/write calls. The fact that you can
           get away with it only on some stacks implies that there is
           no long-term guarantee. Specify -DCAN_WRITE_BEFORE_SENDING_FD
           if you feel brave. File: util/sys_defs.h.
    
           Robustness: re-compile all object files after the "make
           makefiles" options have changed. Files: src/*/Makefile.in.
    
    20050417
    
           Safety: don't call syslog from a user-triggered signal
           handler, and other minor fixes back-ported from the 2.3
           development release. File: postdrop/postdrop.c.
    
    20050422
    
           Bugfix: postsuper could lose an error message after reporting
           a fatal error. Fix back-ported from the 2.3 development
           release. File: postsuper/postsuper.c.
    
    20050427
    
           Bugfix: don't panic when the fall-back relay can't be used
           because the local MTA is MX for the destination, or when
           the fall-back relay can't be used because it was already
           tried via a cached session. Files: util/argv.c,
           smtp/smtp_connect.c.
    
    20050517
    
           Bugfix: in a DSN report, the original recipient should not
           be xtext encoded. File: bounce/bounce_notify_util.c.
    
    20050523
    
           Bugfix: mymalloc() panic with mistyped server host list.
           File: global/dict_pgsql.c.
    
    20040530
    
           Bugfix: TLS MUST_NOPEERMATCH didn't work (inherited from
           TLS patch), and a dangling pointer in the corresponding
           error handling. File: smtp/smtp_proto.c.
    
    20050615
    
           Cleanup: the SMTP client now sends QUIT when the initial
           HELO handshake fails. it still doesn't send QUIT when the
           server greets with a [45]XX code, as that is handled in the
           connection management code before a session context exists.
           File: smtp/smtp_connect.c.
    
    20050616
    
           Bugfix: missing or mis-placed va_end() macros, found in
           Postfix 2.3 code review. Files: util/netstring.c,
           util/myaddrinfo.c, util/attr_clnt.c, util/vstream.c.
    
    20050621
    
           Portability: file descriptor passing is available for Tru64
           UNIX, but AIX4 and IRIX6 will have to do without. This means
           no SMTP connection caching for those platforms. Albert
           Chin. File: util/sys_defs.h.
    
    20050630
    
           Portability: the connection caching code broke on LP64
           systems (inherited from Stevens Network Programming). Files:
           util/unix_send_fd.c, util/unix_recv_fd.c. This code is
           back-ported from the Postfix 2.3 snapshot release.
    
    20050706
    
           Robustness: the SMTP client now disables connection caching
           when it is unable to communicate with the scache(8) server,
           instead of looping forever and not delivering mail. File:
           global/scache_clnt.c. This code is back-ported from the
           Postfix 2.3 snapshot release.
    
           Portability: after sending a socket, the scache(8) server
           now waits for an ACK from the connection cache client before
           closing the socket that it just sent. Files: scache/scache.c,
           global/scache_clnt.c. This code is back-ported from the
           Postfix 2.3 snapshot release.
    
    20050711-3
    
           Portability: on LP64 systems, integer expressions are int,
           but sizeof() and pointer difference expressions are larger.
           Point fixes for a few discrepancies with variadic functions
           that expect int (the permanent fix is to change the receiving
           modules, but that results in too much change, and is not
           allowed in the stable release). Files: tls/tls_scache.c,
           util/clean_env.c, util/vstring.h, smtpstone/qmqp-source.c.
    
    20050806
    
           Workaround: accept(2) fails with EPROTO when the client
           already disconnected (SunOS 5.5.1). File: sane_accept.c.
    
    20050815
    
           Workaround: old Solaris compilers can't link an archive
           without globally visible symbols. File: tls/tls_misc.c.
    
    20050922
    
           Bugfix: the *SQL clients did not uniformly choose the
           database host from the available pool of servers due to an
           off-by-one error, so that the "last" available server was
           not selected. Leandro Santi. Files: dict_mysql.c, dict_pgsql.c.
    
    20050929
    
           Paranoia: don't ignore garbage in SMTP or LMTP server replies
           when ESMTP command pipelining is turned on. For example,
           after sending ".<CR><LF>QUIT<CR><LF>", Postfix could recognize
           the server's 2XX QUIT reply as a 2XX END-OF-DATA reply after
           garbage, causing mail to be lost. The SMTP and LMTP clients
           now report a remote protocol error and defer delivery.
           Files: smtp/smtp_chat.c, smtp/smtp_trouble.c, lmtp/lmtp_chat.c,
           lmtp/lmtp_trouble.c.
    
    20051011
    
           Bugfix: raise the "policy violation" flag when a client
           request exceeds a concurrency or rate limit. File:
           smtpd/smtpd.c.
    
           Bugfix (cut-and-paste error): don't reply with 421 (too
           many MAIL FROM or RCPT TO commands) when we aren't closing
           the connection. File: smtpd/smtpd.c.
    
    20051013
    
           Bugfix: don't do smtpd_end_of_data_restrictions after the
           transaction failed due to, e.g., a write error. File:
           smtpd/smtpd.c.
    
           Cleanup: the SMTP server now enforces the message_size_limit
           even when the client did not send SIZE information with the
           MAIL FROM command. This protects before-queue content
           filters against over-size messages. File: smtpd/smtpd.c.
    
    20051105
    
           Workaround: the next-hop logical destination information
           for connection caching was reset only after a good non-TLS
           connection, so that cached connections to non-TLS backup
           servers could suck away traffic from TLS primary servers
           (the Postfix SMTP client cannot cache an open TLS connection).
           Found during code review. Fixing this requires more change
           than is allowed in a stable release. File: smtp/smtp_connect.c.
    
    20051108
    
           Bugfix: two messages could get the same message ID due to
           a race condition. This time window was increased when queue
           file creation was postponed from MAIL FROM until the first
           accepted RCPT TO. The window is closed again. Found by
           Victor. Files: global/mail_stream.c, global/mail_queue.c,
           cleanup/cleanup_message.c. This code is back-ported from
           the Postfix 2.3 snapshot release.
    
    20051119
    
           Bugfix: the queue manager did not write a per-recipient
           defer logfile record when the delivery agent crashed after
           the initial handshake with the queue manager, and before
           reporting the delivery status to the queue manager. Files:
           *qmgr/qmgr_deliver.c.
    
    20051126
    
           Log warning when REDIRECT, FILTER, HOLD and DISCARD are
           used in smtpd_etrn_restrictions. File: smtpd/smtpd_check.c.
    
    20051128
    
           Bugfix: moved code around from one place to another to make
           REDIRECT, FILTER, HOLD and DISCARD access(5) table actions
           work in smtpd_end_of_data_restrictions. PREPEND will not
           be fixed; it must be specified before the message content
           is received. Files: smtpd/smtpd.c, smtpd/smtpd_check.c,
           cleanup/cleanup_extracted.c, pickup/pickup.c.
    
    20051201
    
           Bugfix: the LMTP client would reuse a session after negative
           reply to the RSET command (which may happen when client and
           server somehow get out of sync). Problem found by Christian
           Theune. Files: lmtp/lmtp.c, lmtp/lmtp_proto.c.
    
    20051207
    
           Bugfix: race condition in the connection caching protocol,
           causing the SMTP delivery agent to hang after delivering
           mail, while trying to save a connection. Introduced with
           Postfix 2.2.5. Files: scache/scache.c.
    
    20051208
    
           Bugfix: the best_mx_transport, mailbox_transport and
           fallback_transport features did not write a per-recipient
           defer logfile record when the target delivery agent was
           broken. This the analog of queue manager bugfix 20051119.
           Files: global/deliver_pass.c.
    
    20051212
    
           Bugfix: an EHLO I/O error after STARTTLS would be reported
           as a STARTTLS I/O error. File: smtp/smtp_proto.c.
    
    20051213
    
           Bugfix: the *SQL, proxy and LDAP maps were not defined in
           user-land commands such as postqueue. Leandro Santi. File:
           postqueue/postqueue.c.
    
    20051225
    
           Workaround: the regexp map cleverly avoided scanning constant
           results for non-existent $number expressions, but failed
           to subject those constant results to $$ -> $ replacement.
           Since fixing this requires too much change for a stable
           release, the optimization is turned off (it doesn't exist
           in the 2.2 PCRE map, either). Files: util/dict_regexp.c.
    
    20051231
    
           Bugfix: the anvil server would terminate after "max_idle"
           seconds, even when this was less than the anvil_rate_time_unit
           interval. File: anvil/anvil.c.
    
    20060101
    
           Portability: 64-bit support for LINUX chroot script by Keith
           Owens. File: examples/chroot-setup/LINUX2.
    
    20060103
    
           Backout: the Postfix 2.2.6 paranoia about garbage in remote
           server replies caused "multiple delivery" problems or "no
           delivery" problems with broken servers/firewalls. Postfix
           still logs a warning but no longer defers delivery. File:
           smtp/smtp_chat.c.
    
           Portability: FreeBSD 6 is a supported platform. Files:
           util/sys_defs.h, makedefs.
    
    20010604
    
           Safety: new "smtp_cname_overrides_servername" parameter.
           The default value ("yes") is backwards compatible.
    
           With a value of "no", the Postfix SMTP client no longer
           allows CNAME expansion to override the hostname that is
           used for logging, SASL password lookup, TLS policy decisions,
           or TLS certificate verification. Instead it uses the name
           of the recipient domain, the host or domain name specified
           in Postfix configuration files, or the hostnames obtained
           with MX lookups. To prevent cheating with hostnames in MX
           lookup results, you will have to suppress MX lookups with
           explicit [hostname] entries in transport maps. Files:
           dns/dns_lookup.c, dns/dns_rr.c, proto/postconf.proto.
    
    20060108
    
           Bugfix: mailbox_command_maps was not subject to $name
           expansion. File: local/local.c.
    
    20060115
    
           Bugfix: don't ignore the per-site policy when SSL library
           initialization fails. Introduced after adopting the TLS
           patch. File: smtp/smtp_session.c.
    
    20060121
    
           Bugfix: a TLS per-site MUST_NOPEERMATCH policy could not
           override a stronger main.cf policy, while a per-site NONE
           policy could. Fixed with a clean re-implementation from
           Postfix 2.3. File: smtp/smtp_session.c.
    
           Bugfix: a combined TLS per-site (host, recipient) policy
           of (NONE, MAY) changed a global MUST policy into NONE, and
           a global MUST_NOPEERMATCH into MAY. The result is now NONE.
           Problem found by exhaustive simulation. Fixed with a clean
           re-implementation from Postfix 2.3. File: smtp/smtp_session.c.
    
    20060130
    
           Bugfix: an empty remote_header_rewrite_domain value caused
           trivial-rewrite to dereference a null pointer, but only in
           regression tests, not in production. Postfix rewrites
           addresses in the remote rewriting context only when the
           remote_header_rewrite_domain parameter value is non-empty.
           File: trivial-rewrite/rewrite.c.
    
    20060202
    
           Workaround: a malformed domain name lookup result (such as
           null MX record) is now treated as a hard error, so that
           Postfix will no longer repeatedly try to deliver mail until
           the message expires in the queue. However, this will not
           reject mail with reject_unknown_sender/recipient_domain.
           That would require too much change for a stable release.
           File: dns/dns_lookup.c.

--
        Viktor.

Disclaimer: off-list followups get on-list replies or get ignored.
Please do not ignore the "Reply-To" header.

To unsubscribe from the postfix-users list, visit
http://www.postfix.org/lists.html or click the link below:
<mailto:majordomopostfix.org?body=unsubscribe%20postfix-users>

If my response solves your problem, the best way to thank me is to not
send an "it worked, thanks" follow-up. If you must respond, please put
"It worked, thanks" in the "Subject" so I can delete these quickly.