Hi Stuart,
thanks for reviewing and feedback, generally your fixed version
still works as it should, just a few comments inline.
On Sun, Jan 18, 2026 at 3:56 PM Stuart Henderson <stu@spacehopper.org> wrote:
On 2026/01/17 17:20, Sebastian Reitenbach wrote:
commentary inline, updated diff below
> C++ Library: stdc++
alarm bells here, libstdc++ is wrong for amd64
| -V= 2016-07-R1
| -DISTNAME= kismet-${V}
| -PKGNAME= kismet-${V:S,-,,g}
| +V= 2025-09-R1
| +DISTNAME= kismet-${V}
| +PKGNAME= kismet-${V:S,-,,g}
| +CATEGORIES= net security
| CATEGORIES= net security
| -REVISION= 1
|
| -HOMEPAGE= https://www.kismetwireless.net/
| -SITES= ${HOMEPAGE}code/
| -EXTRACT_SUFX= .tar.xz
| +HOMEPAGE= https://www.kismetwireless.net/
| +SITES= ${HOMEPAGE}code/
| +EXTRACT_SUFX= .tar.xz
these replace correct tabs with spaces, should go back to tabs
CATEGORIES is doubled
| -WANTLIB+= c m panel pcap ${COMPILER_LIBCXX} curses util pcre
| -LIB_DEPENDS= devel/pcre
| -RUN_DEPENDS= net/wireshark,-text
| +WANTLIB+= c m pcap ${COMPILER_LIBCXX} util
| +WANTLIB += crypto rtlsdr
| +WANTLIB += pcre2-8 sqlite3 ssl
| +WANTLIB += websockets z atomic usb-1.0
| +WANTLIB += mosquitto
WANTLIB could do with being more neat
libatomic is for GCC, this needs patching (I opted to disable by default
in autoconf and add it for ports-gcc builds; might actually only be
needed on a subset of archs though)
| +LIB_DEPENDS= comms/rtl-sdr \
| + lang/gcc/8,-libs \
gcc libs should not be listed directly. if it's needed e.g. on ports-gcc
archs it should come from the gcc module.
| + net/mosquitto \
| + www/libwebsockets
| +RUN_DEPENDS= comms/rtl_433 \
| + devel/py-protobuf${MODPY_FLAVOR} \
MODPY_FLAVOR is no more
The py-* above and below py-websockets, are a left-over from my former
attempt to upgrade Kismet to a git checkout about 2 years ago. Meanwhile
the python modules are all deprecated, and not used anymore, so I removed both py-*
completely.
| + net/wireshark,-text \
| + net/py-websockets
|
| .include <bsd.port.arch.mk>
|
| -post-configure:
| - ${SUBST_CMD} ${WRKSRC}/conf/kismet.conf \
| - ${WRKSRC}/man/kismet_drone.conf.5
| - cd ${WRKSRC} && ${GMAKE} depend
| -
| post-install:
| ${INSTALL_DATA_DIR} ${PREFIX}/share/doc/kismet
| - ${INSTALL_DATA} ${WRKSRC}/README \
| + ${INSTALL_DATA} ${WRKSRC}/README.SSL \
| ${PREFIX}/share/doc/kismet
have left for now but I don't think that file is really worth
installing
| @cd ${WRKSRC} && ${GMAKE} binsuidinstall
set INSTALL_TARGET instead
| +- // rtl_433 -F json -M level -d X -f Y [additional]
| +- unsigned int num_args = 9;
| ++ // rtl_433 -F json -M level -s 1024k -d X -f Y [additional]
| ++ unsigned int num_args = 11;
| +
| + // Channel, if any
| + char *channel = NULL;
| +@@ -452,6 +452,8 @@ int open_callback(kis_capture_handler_t *caph, uint32_
| + local433->rtl_argv[n++] = strdup("json");
| + local433->rtl_argv[n++] = strdup("-M");
| + local433->rtl_argv[n++] = strdup("level");
| ++ local433->rtl_argv[n++] = strdup("-s");
| ++ local433->rtl_argv[n++] = strdup("1024k");
I haven't touched this but would it be better to patch in rtl_433?
I preferred to not touch rtl_433, because it provides an option to adjust sampling
frequency.
Actually all Kismet patches are already merged upstream. This particular, upstream maintainer
is watching, if that lowered sample rate causes issues on other OSs. If that's the case,
he'll make the sample rate a configurable option.
| +lib/pkgconfig/kismet.pc
I haven't touched this yet, but VERSION is broken, they forgot to run
tools/mkversion.sh before rolling release tarballs
upstream maintainer is thinking about maybe releasing a new version in a couple
of weeks. I'll point that out to him.
| --- /dev/null 1 Jan 1970 00:00:00 -0000
| +++ pkg/README 17 Jan 2026 16:14:55 -0000
remember to regen PLIST after adding a readme
updated diff below, fixing the above plus a few other little bits, does
this still work for you?
Yes, still all works like a charm wifi, sdr rtl433, sdr_adsb, nrf51822.
So, with your fixed version minus the py-* depends, then OK to update (after importing rtl_433)?
cheers,
Sebastian
No comments:
Post a Comment