bgpdump/ChangeLog

==========================================================
The current maintainer of this library is the
RIPE NCC information services department.

You can reach us at https://bitbucket.org/ripencc/bgpdump

PLEASE DO NOT E-MAIL INDIVIDUAL DEVELOPERS ABOUT
ISSUES, AS YOUR E-MAIL MAY BE LOST
==========================================================
2015-10-08 Alexis Fasquel <alexis@pch.net>
	* Adding support for Extended Timestamp MRT Header corresponding to MRT type BGP4MP_ET
	  Note that this changes the semantics of the output of bgpdump - when a messages
	  with Extended Timestamp MRT Header is parsed, the microseconds attribute
	  is appended to the timestamp (.xxxxxx)

2015-09-19 Colin Petrie <cpetrie@ripe.net>
	* Adding support for BGP4MP_MESSAGE_LOCAL and BGP4MP_MESSAGE_AS4_LOCAL subtypes
	  Note that this changes the semantics of the output of bgpdump - when an entry is
	  prefixed by BG4MP_LOCAL (-m/-M) then the output fields are the *destination* IP/AS
	  rather than the source.

2015-07-16 Max Lapan <mlapan@ripe.net>
	* -p option to show packet index

2011-01-25 Devin Bayer <dbayer@ripe.net>
	* move development to bitbucket

2010-10-21 Devin Bayer <dbayer@ripe.net> v1.4.99.13
	* fix configure script for Darwin

2010-09-24 Devin Bayer <dbayer@ripe.net> v1.4.99.12
	* Remove --disable-ipv6 flag
        * Fix parsing of IPv6 NEXT_HOP attributes
        * Add clearer command-line options
        * Add logging to STDERR
        * Add dumping to files instead of STDOUT
        * Add regression test suite

2009-11-24 Roman Kalyakin <roman@ripe.net> v1.4.99.11
	* Fixed incorrect output when running with -m which was
	  introduced by a change in one of the previous versions.
	  (issue IS-439)

2009-10-13 Ruben van Staveren <ruben@ripe.net>
	* Fix segfaults with empty AS path on -m/-M 
	  (issue IS-292)
	  Patch contributed by Lorenzo Colitti <lorenzo@google.com> 

2009-07-29 Erik Romijn <eromijn@ripe.net>
	* Fixing incorrect as path length count in case of AS_(CONFED_)SEQUENCE
	  (issue IS-141)
	  Patch contributed by Toby Ehrenkranz <tehrenkr@cs.uoregon.edu>
	* Added support for the ORIGINATOR_ID and CLUSTER_LIST attributes
	  (issue IS-181)
	  Patch contributed by Maurizio Pizzonia <pizzonia@dia.uniroma3.it>
	* Fixed signedness of MED and localpref (issue IS-180)
	  Patch contributed by Steve Hill <cheesy@sackheads.org>
	* Implemented conditional checking for compression libraries (issue IS-179)
	  Patch contributed by Steve Hill <cheesy@sackheads.org>

2009-07-27 Ruben van Staveren <ruben@ripe.net>
	* Enabled extra compiler warnings
	* Fixed some potentially unsafe constructs (pointer arithmetic,
	  signed/unsigned comparisons, shadowed variables) 

2009-07-22 Ruben van Staveren <ruben@ripe.net>
	* Fixed a case where during -M dumps aspaths where printed for entries that
	  did not have their BGP_ATTR_AS_PATH flag set. Such entries will now print
	  an empty field. Reported by many people.
	* Changed -fpic to -fPIC: newer gccs coerce -fpic to -fPIC any way.
	  difference is that with -fpic the global offset table has a maximum size
	* Fixed nit in getopt handling
	* Added distclean target
	* In bgpdump_free_mp_info(), set pointers to NULL after freeing them.
	  Reported by Robert Kisteleki <robertk@ripe.net>
	* Default output for 32bit ASN numbers is now ASPLAIN

2008-02-01 Erik Romijn <eromijn@ripe.net>
	* Fixed a few memleaks based on a patch from
	  Antony Antony <antony@phenome.org>
	  (Ticket NCC#2007064868)
	* Fixed compiling with --disable-ipv6 based on a patch
	  from Haniff Rahbari <hnrahbari@yahoo.com>
	  (Ticket NCC#2007111325)
	* Fixed a few warnings, mostly based on a patch from
	  Bruno Quoitin <bruno.quoitin@uclouvain.be>
	  (Ticket NCC#2007123069)
	* Made bgpdump's usage message nicer

2007-06-21 Erik Romijn <eromijn@ripe.net>
    * Fixed incorrect dump time display for some cases

2007-06-06 Erik Romijn <eromijn@ripe.net>
    * Set proper compile options to work on Mac OS X 10.4

2007-06-01 Erik Romijn <eromijn@ripe.net>
    * Fixed several bugs in 64-bit PC support and added
	  bzip2 support, credit goes to
	  Bernhard Tellenbach <betellen@tik.ee.ethz.ch>
	  for providing patches.
	* Fixed incorrect IPv6 nexthops in -M/-m output

2007-03-29 Erik Romijn <eromijn@ripe.net>
	* Fixed a serious segfault in the insertion of update dumps

2007-03-28 Erik Romijn <eromijn@ripe.net>
	* Fixed a large memory leak in TABLE_DUMP_V2 processing

2007-03-27 Erik Romijn <eromijn@ripe.net>
	* Fixed a bug causing segfaults on processing some MP_REACH_NLRI's

2007-03-09 Erik Romijn <eromijn@ripe.net>
	* Updated TABLE_DUMP_V2 support to comply with draft-ietf-grow-mrt-04

2007-03-07 Erik Romijn <eromijn@ripe.net>
	* Built in TABLE_DUMP_V2 support
	* Rename ..MESSAGE32 to ..MESSAGE_AS4
	* Fixed bug where -M always crashed

2007-02-01 Erik Romijn <eromijn@ripe.net>
	* Minor fix in ASN32 support

2006-10-20 Lorenzo Colitti <lorenzo@ripe.net>
	* Implement ASN32 support

2005-11-08 Young Hyun <youngh@caida.org>
	* IPv6 prefixes were not printed properly by bgpdump.

2004-02-10 Lorenzo Colitti <lorenzo@ripe.net>
	* Fix a stupid crash when reading truncated prefixes.

2004-02-09 Lorenzo Colitti <lorenzo@ripe.net>
	* Decode unknown attributes properly.
	* Fix a crash when reading truncated prefixes.

2004-02-04 Hong-wei Kong  <hong-wei_kong@agilent.com>
        * Add a new option (-t dump|change) to bgpdump. If '-t dump' is 
	  specified, then the decoding result with '-m' / '-M' option will
	  show the timestamp when these routes are dumped. If '-t change' is
	  specified, then the decoding result with '-m' / '-M' option will
	  show the timestamp when the routes were last time changed. This 
	  option only works with routing table dumps.

2004-01-16 Lorenzo Colitti <lorenzo@ripe.net>
	* Fix compile errors under cygwin (only IPv4 works though)

2003-12-18 Hong-Wei Kong <hong-wei_kong@agilent.com>
	* Added support for dumping BGP routing table entries in
	  machine-readable format (-m). Previously -m was only supported
	  for BGP messages

2003-12-16 Shufu Mao <msf98@mails.tsinghua.edu.cn>
	* Revised the decoding of BGP routing table dump to make the result
	  identical to route_btoa

2003-12-15  Lorenzo Colitti <lorenzo@ripe.net>
	* Proper autoconf support: check for features instead of target
	  systems, remove extra .h files and only use configure.in, use
	  Makefile.in instead of Make.include, link both static and
	  shared library, check for ar and ranlib, etc.
	* Changed INET6 to BGPDUMP_HAVE_IPV6
	  (N.B. Docs are not updated yet)
	* Fixed possible endianness issues on Sparc.
	* Fixed a crash while decoding NOTIFICATION messages

2003-11-26  Shufu Mao <msf98@mails.tsinghua.edu.cn>

	* Workaround for zebra "NOIP" bug
	* Preliminary autoconf support (./configure; make)
	* Added option to disable IPv6. If this option is specified to
	  ./configure, the library is compiled without IPv6 support.
	  The default is to turn on IPv6 if the necessary headers are
	  available.

2003-11-25  Lorenzo Colitti <lorenzo@ripe.net>

	* Added support for OPEN and NOTIFICATION messages.
	* Better workarounds for corrupt data produced by zebra/quagga bugs.
	* The library now keeps track of how many records were parsed and how
	  many were successfully parsed.
	* Better error reporting. Now an error will be logged through syslog
	  whenever a record cannot be decoded.
	* Agilent Labs contributed bgpdump, a tool which emulates route_btoa
	  using libbgpdump. This should be a drop-in replacement for people
	  currently using route_btoa and supports IPv6 on all platforms.

2003-10-15  Lorenzo Colitti <lorenzo@ripe.net>

	* 1.3.1 beta released
	* Added support for reading from gzipped input files and gzipped stdin.
	  NOTE: this means you must now link with zlib!
	* Changed Makefile to build a shared library by default.
	* Renamed test program from test2 to testbgpdump.

2003-10-15  Lorenzo Colitti <lorenzo@ripe.net>

	* 1.3 beta released
	* Added IPv6 support. Note that this changes some interfaces!
	  See README.ipv6 for details.
	* Set announce and withdraw arrays to NULL to avoid segfaults if there
	  is a parse error.
	* Change mstream behaviour so it always returns zero bytes if reading
	  past the end of a packet. This eliminates non-deterministic
	  behaviour (random segfaults, infinite loops, ...) when parsing
	  corrupt data, including the consequences of the zebra "update packet
	  truncated at 4096 bytes" bug.
	* Other fixes to handle corrupt data better.
	* Other miscellaneous fixes.