* buildlink bin/libnet{10,11}-config to bin/libnet-config for comvenience
(they ware renamed in pkgsrc to avoid conflict)
* remove -lnet from BUILDLINK_CPPFLAGS.libnet11.
linker flags should not be in CPPFLAGS, and it berak likage with libtool
as reported in PR 37300.
* libnet11 install just a static library, so set defaut DEPMETHOD = build
Bump PKGREVISION for libnet11 related packages (net/isic will be updated later).
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
- Drop maintainership
- Use an option group for libnet version
- Add USE_DESTDIR support
v1.23 Feb 23 2008
- fixed remotely triggerable NULL dereference in ip_fragment.c
- fix DLT_PRISM_HEADER linkoffset calculation
- check for DATA_FRAME_IS_QOS in wireless frames
- free queued tcp segments with too old seq
v1.22 Jul 22 2007
- in TCP stream, the byte with absolute offset 0 was treated as urgent data;
fixed
- DLT_IEEE802_11_RADIO handling
- added a few missing checks for failed malloc
v1.21 May 10 2006
- more externals to access libnids' intrinsics from the outside
- nids_unregister_*()
- UDP checksumming fix (0 is not an error according to RFC768)
- nids_params.tcp_workarounds
- nids_params.multiproc and queue_limit: merged a patch which creates a
separate thread for packet capture;
- in killtcp.c, send two more RST packets (required because of MS05-019
patch)
- glibc 2.4 syslog.h disaster workaround
and add a new helper target and script, "show-buildlink3", that outputs
a listing of the buildlink3.mk files included as well as the depth at
which they are included.
For example, "make show-buildlink3" in fonts/Xft2 displays:
zlib
fontconfig
iconv
zlib
freetype2
expat
freetype2
Xrender
renderproto
RECOMMENDED is removed. It becomes ABI_DEPENDS.
BUILDLINK_RECOMMENDED.foo becomes BUILDLINK_ABI_DEPENDS.foo.
BUILDLINK_DEPENDS.foo becomes BUILDLINK_API_DEPENDS.foo.
BUILDLINK_DEPENDS does not change.
IGNORE_RECOMMENDED (which defaulted to "no") becomes USE_ABI_DEPENDS
which defaults to "yes".
Added to obsolete.mk checking for IGNORE_RECOMMENDED.
I did not manually go through and fix any aesthetic tab/spacing issues.
I have tested the above patch on DragonFly building and packaging
subversion and pkglint and their many dependencies.
I have also tested USE_ABI_DEPENDS=no on my NetBSD workstation (where I
have used IGNORE_RECOMMENDED for a long time). I have been an active user
of IGNORE_RECOMMENDED since it was available.
As suggested, I removed the documentation sentences suggesting bumping for
"security" issues.
As discussed on tech-pkg.
I will commit to revbump, pkglint, pkg_install, createbuildlink separately.
Note that if you use wip, it will fail! I will commit to pkgsrc-wip
later (within day).
v1.20 Feb 4 2005
- added wscale option parsing; surprisingly, it seems to be in some use
- added nids_dispatch(), for systems which do not ignore pcap timeout
- ability to specify hosts/networks for which we do not check checksums
v1.19 Aug 08 2004
- fixed signed/unsigned comparisons; 1.18 could be possibly crashed in tcp
options parsing (though an unlikely to happen memory layout is required);
now the source is compiled with -W -Wall
- export pcap header of the last received packet (to get timestamp etc)
- export the timeout parameter to pcap_open_live in params
- support DLT_PRISM_HEADER
- support DLT_PPP_SERIAL
- let through dataless acks
- fixed raw_init() prototype
- switched to use %edi instead of %ebx in csum_partial to make gcc-3.5 happy
when compiling with -fPIC; cleaned inline asm
- fixed a bug when a queued FIN segment was not processed properly, which
resulted in not closing a stream
verision of libnet <= 1.0.1b. This will prevent the case where the user
has installed the libnet 1.1.x branch and then tries to install an application
that is not compatible with the 1.1.x tree.
Over time the list of these applications that require the 1.0.x branch
will be reduced as they are updated to later versions that support the
libnet 1.1.x branch.
This addresses PR# 29056 opened by diro (at) nixsys.bz, thanks for the PR !
in the process. (More information on tech-pkg.)
Bump PKGREVISION and BUILDLINK_DEPENDS of all packages using libtool and
installing .la files.
Bump PKGREVISION (only) of all packages depending directly on the above
via a buildlink3 include.
fragment reassembly code. CVE reference for that vulnerability is
CAN-2003-0850.
Note homepage relocation, and take over maintainership since that's
apparently what I've become.
v1.18 Oct 15 2003
- reject tcp packets with old timestamp; needed to pass fragroute test;
well, linux 2.0.36 did not support this ;)
- fixed memory corruption which could be caused by overlarge TCP packets
- adjusted checksum.c to not use multiline literals (for gcc 3.3)
- in configure.in, even if found libnet files, try compilation; there is
another library with the same name
- fix a bug in "collect" field handling; if you did collect-- and then
collect++ (which is rare), you would get a single junk packet
- correct handling of exec_prefix in configure.in
- unlink config.status in "make distclean"
- use pcap_hdr->caplen instead pcap_hdr->len; the only gain seems to be to
gracefully handle pcap files with too short snaplen
- changed soname to libnids.so.1.x, as binary compatibility is not
guaranteed
- switched to sourceforge as homepage
v1.17 Dec 12 2002
- fixed a stupid bug in TCP reassembly; having received a particular order
of TCP out of frame segments, libnids could lost track of the current
seq, and miss the following data stream
- DLT_FDDI
- benign typo in hash.c
- mentioned usefulness of two process buffering on a fast network
v1.17rc1 Aug 30 2002
- support for libnet-1.1 and --with-libnet=no
- added support for libpcap save files
- finally, DLT_LINUX_SLL is recognized
- removed a horrible assumption on sizeof(pointer); it could result in
segfault in scan.c
- --enable-shared
- __i386 -> __i386__ || __i386 :(
- support for 802.1Q VLAN
- support for wireless frames (DLT_IEEE802_11)
- got rid of (obsolete) pcap_open_live_new
- bail out if link type is unknown, instead of pretending it is ethernet
- $(MAKE) -> $(MAKE) $(AM_MAKEFLAGS)
- added a working link to Ptacek-Newsham paper
- %hi -> %hu :)
- align IP header if necessary (should not be)
- improved libraries detection
- mentioned usefulness od setsockopt(...SO_RCVBUF...) on a fast network
v1.16 Nov 3 2000
- nah, at least a release forced by a security bug. A typo in libnids.c
could cause libnids to segfault when source routed frame has been received.
v1.15 Oct 9 2000
- token ring support
- new configurable option (non-default): if a tcp callback hasn't processed
all available data, it is called immediately again
- fixed alignment in hash.c, which caused sigsegv on Sparc
- another _obviously_ redundant include file added to configure test progs
- html version of the API documentation
foo-* to foo-[0-9]*. This is to cause the dependencies to match only the
packages whose base package name is "foo", and not those named "foo-bar".
A concrete example is p5-Net-* matching p5-Net-DNS as well as p5-Net. Also
change dependency examples in Packages.txt to reflect this.
