Changelog for kismet-git-0.0.0+git.20190102-lp150.1.1.x86_64.rpm :

* Thu Jan 03 2019 mardnhAATTgmx.de- Update to version 0.0.0+git.20190102:
* Don\'t use WPS manuf as ieee manuf; this also fixes a problem clobbering the global \'unknown\' manuf
* Handle client WPS data
* Stop gcc from whining (clang is fine)
* Increase precision for float-to-int for averaging
* Switch to logging blob instead of text for json, but should be the same output
* copy/pasta mistake fixed
* Compiled: Added Tello and Osmo
* Just added a couple, Tello and Osmo
* Sun Dec 30 2018 mardnhAATTgmx.de- Update to version kismet201808beta1+git.20181227:
* Copypasta w.r.t. # of characters in rtladsb
* Tweak type names for simplicity
* Fix channel hop rate calculations which were done completely wrong
* Fixed adsb and amr bugs
* Switch to $(INSTALL) and using the env variable
* Update API
* Add KISMET_ETC environment variable for helper scripts
* Modern API
* Trim temp files
* Exclude some temp files
* Removed oopsie files
* Added rtlamr and rtladsb support
* Add cisco client mfp
* Fix MFP byte selector
* Copypasta
* Add MFP client 221 tag parsing
* Add ie150 ccx tag processing Mildly revamp how tags are calculated by rewinding the stream for code readibility
* Support v1 and v2 of QBSS reports
* Add unsigned long long to sqlite_as templates
* Add initial parsers for vendor ie150 tags and cisco 150 txpower tags
* Clean up kismet_server code slightly Remove ncurses dependency and inject a top margin with ansi codes directly
* Fix sorting of the UI details
* Include field ID in assignment errors
* Return 200 OK for drop endpoint success
* Add packet drop endpoint
* Fix _DELETE sql wrapper; it doesn\'t take fields
* Add _DELETE query and generic-ops to query
* Update view for formatting and consistency
* Fixed url typo
* Junk old files
* Update docs
* Added docs as submodule Added submoduel directions
* Move docs in prep to pull submodule
* Fix regex on views
* Imported new jekyll-templated version of the readme and developer docs; these are slightly more annoying to deal with in the git repository but fantastically better for dealing with on the web site. Updated readme to point to the new split apart docs tree and the generated website tree
* return error on kismetdb pcap when kismetdb not enabled
* Standardization of API names, by-time to last-time
* use SQLITE_TRANSIENT on all string bindings
* Fix transient strings in sqlite3
* Disable/remove \'channel\' option since the kismetdb log doesn\'t currently hold a channel record
* 404 if dblogging is not enabled Add GET pcapng kismetdb endpoint
* Convenience accessors for cached variables
* Check for auth on dot11 pcap and kismetdb pcap
* Allow for pausing buffer fill to populate from a database while the buffer flushes Enable the new historic pcap api
* Fix initializing statement, fix wrapping fields in ()
* More robust startup checking in kismet server Modernize some of the alertracker code Fix alertracker aborting early and causing unitialized class variables when config is wrong
* Add new global tag
* Update API usage for globals Exit on fatal conditions properly instead of allowing initialization to continue to the pollable section, this properly exits much earlier for config file errors
* Add disable for nm and lmsensors; disabling nm in particular is not recommended
* Expand sqlite_cpp11 api slightly, fix linkage errors Convert stub of the packet history api to new sqlite api Tweak simple post endpoints to include the URI Rename packet history api from kismetlog to kismetdb Add LIMIT to kismetdb api
* Enable new code
* Start adding simpler c++11-idiomized sqlite interface
* Expand pythonrest module to include new clients-of and access-point endpoints
* Update docs
* Add device_list_by_mac(...) to rest connector
* Remove pass-by-referece scope locking since it breaks some GCCs, it seems
* Trim some cruft from the autoconf process to make it marginally faster
* Bulk revert boost::geometry usage, it\'s not reasonable without pulling in an insane amount of other boost code
* Continued expansion of boost header-only inclusion
* Add boost MPL lib required by boost geom
* Add poylgon geometry
* Allow initializing a trackerelementvectordouble from a std::vector
* Enable building new geometry
* Move point to new file Start a boost::geometry based set of tracked data
* Rename namespace for consistency
* Add /devices/multimac/devices.json endpoint
* Re-order code slightly and simplify API
* small fix to the disable libusb command line help
* Add 2d point
* Alphabetize library credits Add boost:geom
* Add the header-only boost::geometry library
* minor consistency tweak
* Fix string search by mac
* Add /phy/phy80211/clients-of API to extract client devices in a single request
* Add operator>> to device_key for templated transforms
* Add toggle to turn off AP view Add docs in memory.conf Add proberesp ap type Add proberesp and adhoc to ap view
* Fix simple endpoint generator functions
* Add directory detection to alert about pw dir
* Minor docs
* Clean up some copypasta
* Tweak doc formatting
* Docs
* Docs
* Fix 404 on POSTs Enhance 404 error page
* typo in path
* Revamp device summarization to return a sane value and auto-process vectors Add time-based endpoint for views Add view docs Simplify (ha) the simple api to trim the empty [0]th token off a URI
* Set stream error bits when processing uuid and mac via operator>>
* Add generic string transform using templates
* Start enablng first view
* Start enabling view engine
* Split workers definition for consistency
* Near-complete implementation of view rest api Tweaks to simplified rest api to include uri Migration of uri detection to kismet_http namespace
* Tweak new API
* Expand simplepost api to include original post data
* Fix typo in devicetracker endpoints Work on common endpoint for views / datatables
* Populate regex, stringmatch workers
* Rename for consistecy
* Remove migrated defs
* Start splitting workers
* Roll back all of the python3-specific changes because protobuf doesn\'t work with py3
* Silence junk errors from probe and list phases of interface building
* More work on views
* More incremental work on the devicetracker view manager
* Simplify callbacks with bind vs lambda stubs
* Add mutex-locking proxy macro
* dot11 fingerprint comparison
* Add fingerprint saving
* Remove const from dynamic template sets
* Add generic asString methods
* Explicitly convert to string for py3 (in theory)
* Remove some debug info
* More modernization of the configfile engine
* Update error for python3-setuptools
* Add operator>> to uuid for generic conversion
* Add operator>> for macaddr to load from generics
* Start updating config file for modern C++, simplify several options
* More work on fingerprint and docs for dot11 fp
* More work on dot11 fingerprinting API
* Provide empty map when no JSON provided in POST
* Lock muted during complex path callbacks
* Add additional http endpoint handlers
* Tweak config file api slightly
* Remove globalreg direct reference in configfile
* Use kis endian shim
* Use __APPLE__ instead of __SYS_DARWIN__
* Add \'auth\' option to simple tracked and simple post endpoints
* Start adding dot11 fingerprint api
* Add a simple POST handler w/ callback function to simplify the HTTP engine API further
* Remove bogus exec bits
* Add some error handling to external
* Fix autoconf to look for python3 Fix capture tool makefiles to use python3 Manually fix protobuf not building valid python3 relative includes
* Start updating packaging for python3
* Migrate to python3 - it\'s getting harder to automatically support python2 on many distributions and it\'s EOLing in the forseeable future.
* Simplify fingerprint hash to 32bit xxhash32 per category
* Rename beacon and probe fingerprint keywords
* Fix malformed fmt::format call for crypto change
* Split fingerprint
* Change to using the https://linuxnet.ca/ieee/oui/ sanitized OUI list, regenerate manuf file
* Sync manuf with IEEE
* Add chrono to util
* Move kismet_manuf.txt to $SHARE/kismet/kismet_manuf.txt to get it out of etc.
* Revamp macaddr processing somewhat, update some methods Remove slow, broken macmap construct and just use std::map with proper macaddr masking Remove old, un-ported filter system for prep to new more powerful filtering in the future
* Fix copypasta for forceconfigs of the new 80211 config
* Show fingerprint data in UI; first revision
* Fixed sth. confused
* Switched to standard xxhash32 library for endian safe-ness Created own C++ wrapper for xxhash32 Migrated to xxhash32 canonical output Fixed hashing of integer values and endian-ness
* Update all channel sizes to the proper size_t instead of unsigned int
* Add logging info to readme
* Make errors match config parameters
* Thu Oct 25 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1540475259.b3d06871:
* Merge patch from Steev to fix mispositioned packed attributes
* Actually make osx building happy again
* Revert \"Add unsigned long long to try to make OSX happy\"
* Add unsigned long long to try to make OSX happy
* Started adding fingerprinting for client devices
* Start tweaking fingerprint more
* Tweak headers slightly to try to resovle mojave issue
* Fix mis-hashing of nonexistent tags in fingerprints
* Switch to xxhash32 instead of std::hash for more consistent tag hashing
* Start defining probe fingerprinting
* Use tuple-based IE 221 tag counting in the device fingerprinting system
* Start adding IE tag fingerprinting config options
* Detect the new variant of the 8812/8814 drivers
* Add new config to packaging
* Add 802.11 config Start adding dot11 fingerprinting as a conf option
* Start adding access point finger printing in prep for whitelisting/baselining
* Remove ill-planned refinement functions
* Simplify uuid rets
* Added fedora package list from David Poole
* API tweaks
* Apply partial sort to the refinement functions
* Start using partial_sort for the main device lists to get a little CPU bonus
* Shim partial sort for parallel
* Start fixing up refinement function
* Add std::hash-compatible map primitive (size_t key)
* Add a boost-derivative header
* Update kismet_log_to_kml.py
* Add tag list to probessid
* Track advertised IE tags, in order, in SSID records
* Allow set of vectors from iterators
* Start adding capabilities tags
* json serializing vectorstring
* Add vector of strings as primitive
* Expand parsing of VHT capabilities
* Add ie 70 RM measurement tag framework Add ie 127 extended capabilities tag framework
* try again, moving all constexpr to c++17 only because older GCC still complains
* Fix some missing constexpr Move all constexpr statements to C++14 to deal w/ really old compilers
* Fix missing http server bind in pcap
* Add const and constexpr decorations to tag parsers
* Fix typo and add clarification for client supported channels Add cryptset to probe records
* Add some UI for min/max tx power and channel list
* COpypasta field names
* Fix bug in ie33 Start processing ie33 and ie66 into client device records
* Add dot11 ie36 supported chanels
* copypasta
* Add ie33 probe req power parser
* Make sessioncache path a KismetConnector argument
* remove pcap headers
* Completely revamp janky handshake pcap writer to just manually write a pcap stream instead of using libpcap
* Fix handshake pcap causing a crash; in doing so, discovered that the entire webserver registration process has a lurking bug linked to C++ parent instantiation order. Move http server binding out of HttpHandler parent class and into each HTTP final object.
* Include driver type in kismet log report
* Fix botched commit
* use fields list
* Fix handling of zero-length summarization lists
* Remove legacy python stuff
* Add mutex locking to simple endpoints to prevent introducing a race condition that may have been crashing during system_monitor
* More work on refinement generic; rename worker functions for consistency
* Array slicing construtors
* More work on the refinement function
* Further tweaks to pcre worker and refine parameters
* Add conversion function to vector of pairs
* Start refactoring to split out device windowing
* Trim unused functions from util.cc, modernize other functions for const string refs, optimize other functions for speed, minimize std::string append-building behavior
* Add descriptions to packages
* Rework support for nexmon now that it handles netlink
* Update versioning
* Handle probe and list more gracefully if the rtl433 binary isn\'t present
* Minor doc updates
* Revamp manufacturer tracking to track a string object instead of duplicating strings to every device
* Added manuf_lookup=true/false to kismet_memory.conf to optionally turn off manuf mapping to save ram
* Fix up uav compiling script Add script to fetch IEEE manuf and generate a kismet manuf file Add a kismet manuf file instead of counting on wireshark to be installed Revamp manuf parsing to allow for longer manufacturer names, cleaner parsing Allow formatted directories in manuf files
* Manually populate systemd
* Add systemd to packaging Fix non-debug packaging Add running-as-root warning Add systemd config docs
* More incremental work on db streaming
* Update streamtracker to globalreg nameless api
* Start implementing filtering and the databaselogfile rest URI
* Create dynamic binding API for sql queries
* Additional doc updates
* Minor docs updates
* Slight update for new API
* Plumb POST to kis external
* Fix populating http fields Extend proxy demo to use http GET fields
* Rename the kismet external api http POST data field to just variable_data Send http GET vars with the external http api call
* Naming consistency
* Clean up fpm scripts
* Re-order autoconf to not probe for all C++ standards if it finds the newest
* Handle limit-packets and outfile properly in pcap Detect missing source file in all log processors
* add friendlier errors to the data source control URIs
* Add logtools to recommended packages
* Add log tools package and metapackage
* python3 print() updates Friendly package errors
* Fix print() for python3 Add friendly failure for missing deps
* +x csv log
* Flag replacements on packages
* Tweak packages for xenial
* More fixes for exception catching
* | is not ||; fix nullptr compare
* Adding Kismet Log to CSV tool.
* More exception references
* Split debug package to its own version name
* Fix stupid error in new helper probe
* More exception cleanup
* Use references for exceptions
* Flag all capture packages as recommended installs on core
* Gracefully handle IPC programs not present (req\'d for splitting capture tool packages up)
* Add function to see if external binary is available in prep for rewriting how datasources probe
* Fix null reference in nrf when getting packets w/out a packinfo
* Fix http data dir
* Fix conf fpm packaging
* catch exception by reference
* Tweak scripts to be run from main dir
* Update NRF package gen
* Add debian control files for use with FPM package generation Update package gen
* More work towards packaging standards
* Use python install to install the scripts
* Start adding FPM based package creation
* More updates to deb packaging
* Start updating deb package
* Log error values from GPSD as part of the snapshot
* Migrate UUID output to fmtlib
* Use fmtlib on macaddr for faster mac2strig conversion
* Start adding external antenna mapper
* Migrate system_monitor to new simple endpoint
* Expanded the basid endpoint server to take an optional generator function
* Add simplified endpoint class for basic serving of records
* Tweak some includes in mousejack
* Add nRF docs
* Actually populate mousejack
* Fix silly copypasta Disable channel debug print
* Fix typos / proper install of nrf
* NRF clasifiers
* Indentation
* ignore
* Fix up building nrf_mousejack
* Enable nrf-mousejack building
* Start activatig nrf phy
* Virtual destructor
* static_ptr_cast != static_cast
* Slight modernization to phy_bluetooth
* Clean up NRF interface
* Junk garbage file
* Debug-functional mousejack reading
* more, but non-working, work on mousejack
* More mousejack work
* More functionality for mousejack
* More NRF work
* Start adding channel support
* Minimal additional work on nrf
* More work on mousejack
* Start adding nrf mousejack implementations
* Add libusb checking
* Start adding encryption column
* Remove legacy statealert code entirely Merge bssts alert into phydot11
* Tweak docs
* Increase the timeout before declaring a datasource IPC dead Add staggered bring-up of very large numbers of local datasources
* Migrate shared locks to devicetracker httpd
* Convert some of channel tracker to shared mutex
* Start converting devicetracker and datasourcetracker to shared mutexes
* Improvements to the new mutex
* Include fmt/ostream by default Convert kismet mutex to use shared read mutexes when C++14 is available, otherwise fall back to complete mutexes
* Add mutex to state alert
* re-indent statealert
* Switch to json string sanitization from nlohmann jsonhpp
* Some thread protection on datasource httpd, also segregate sorting of entire data vector and do it on a copy outside of a thread lock
* Spawn sources in their own thread
* Process any tail of the buffer after the IPC process has died, before we decide it was a fatal condition
* Pollable tracker uses a maintenance queue independent of the main vector of pollables already, so there\'s should be no need to lock the main pollable vector while merging and processing it; these also should only be called from the main loop. This keeps the main thread from holding the pollable queue in lock if another thread is adding/removing pollables
* Don\'t try to split sources which can\'t tune
* Start revamping source launching
* Fix variable expansion in packageconf
* Fix populating common name
* Make more fields dynamic in the dot11 object; including WPS state, dot11d, dhcp strings, cdp strings, and eap identities
* Non-lambda dynamic proxy
* Convert device username to dynamic field to squeeze out a little more RAM
* Add dynamic lambda proxy
* Remove phy80211 packet counter (it didn\'t reflect real packet count anymore with multithreaded dissectors anyhow)
* Possible speed boost to sorting elements for the UI under huge loads
* Add list_interfaces API to KismetConnector
* Update idiom for if not x is / if x is not Handle channel parameter in rtl433
* Only link lmsensors when lmsensors is found
* Thu Aug 02 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1533182157.cfd77eca:
* Purge old non-used export of docs
* Only probe librtl when it\'s needed to silence some unnecessary errors
* Start prepping beta release
* Finish migrating readme to MD for nice web export
* More conversion to markdown
* Start converting readme to markdown
* Added basic add_source python
* Reduce copying in the remote driver code
* Update the readme
* Fix incorrect comments
* Remembering my own API is hard; use function worker properly
* Make some more variables atomic
* Convert some values to atomics
* Convert device logger to a functionworker Make some elements in databaselog atomic
* Simple math is hard: Fix boundary detection in chainbuf
* Don\'t show legend on channel coverage
* Mention --no-ncurses
* Fix improper buffer use in devicetracker http POST handling Fix vector cloning in trackedelement Fix exception catching in HTTP POST and HTTP GET stream handling
* add some safety checking
* Use duplicated vectors to prevent thread conflict on resizing immutable devices
* Restore MSG emission
* Protect buffer with mutexes
* Clean up locking
* Fix quirks with optimization Add -fno-omit-frame-pointer protection
* Remove default move constructors Add vector copy constructor Copy immutable vector before doing bulk operations because it might be getting reorganized during a vector expansion
* Revamp the http stream closing flushing
* Revert \"Flush out buffers before killing the generator threads\"
* Refactor how workers are executed
* Flush out buffers before killing the generator threads
* Clean up more threading and cancellation
* Clean up some of the core conditional locker core
* Try to extract owner of lock on mutex timeout
* Always assign a generic signal from rtap even when the antenna numbering is weird
* Don\'t lock in destructor
* Refactor some of how mutex locking works to try to eliminate some variables
* Convert all sqlite mutexes into a single datbase mutex in case sqlite wasn\'t compiled as thread safe by the platform
* Reduce the scope of the mutexes in databaselogfile to ensure we generate all the complex data before we lock the database
* Update to latest radiotap definitions Handle, at the C++ level, multiple antennas per card in the rtap header Start prepping support for VHT and MCS Handle other modern radiotap fields
* Fix building httpproxytest plugin
* de-virtualize the less-than function which should slightly improve the sorting functions
* Disable krack/noncereuse detector by default bc of false positives in most situations
* ifdef\'d constexpr for older compilers
* Update the stdcxx macro detection, detect if we\'re able to do c++14 or c++17 as well
* broken initial impl of some of the pcap api for databaselog
* Start database pcap stream code
* Don\'t do a full re-dump of the devices on exit
* Fix hop rate in Python module
* Enable compile-time optimization by default
* Added some docs re: compiling capture tools only
* Hopefully fix capture-tools-only build trying incorrectly to build deps for the C++ components of the protobuf files
* Start converting globals to simplified method less prone to typos
* Don\'t trigger buffer errors when recoverable
* Clarify comments
* Fix reversed sort
* Copypasta
* Consistently handle sorting when a device is missing a field
* Handle action_noack (or at least, set their type) per https://github.com/kismetwireless/kismet/issues/69
* Tweak locking orders and avoid holding locks on two devices at once in client and eapol categorizers
* Strip more globalreg passing in favor of proper global namespace Remove globalreg passing from http handlers
* Add buffer size to pcapng log Split pcapng buffer logging from packetchain, move all existing code to packetchain version
* Initial lmsensors-based thermal/fan reporting
* Actually link libsensors when its found
* use stringstream instead of std::stod because some broken compilers still don\'t define it properly
* Add libsensors4 lm-sensors library detection
* (blindly) update centos packaging for new directories
* Remove ancient, incorrect man pages
* More last_bssid completion
* Add wifi BSSID column
* Populate last_bssid properly in dot11
* Fix maxrate not being initialized leading to insane values
* Start prepping for RC1
* Set more channel strings
* Fix alphanum sorting
* Set channel string from freq for all devices
* Properly protect logfiles from multithread packetchain
* Protect device list
* Switch blocking conditionals to notify_all Enable multithreaded packet decoding
* Fix typo in gpsgpsd2 thanks to InformalSecurity https://github.com/kismetwireless/kismet/issues/66
* Start rewriting tracked component docs
* Simplify registering entry
* Register the kismet group as a system group
* Clean up proxydynamictrackable code and erase code to catch some errors
* Add fmtlib debug hook
* Clean up sorting
* Simplify channel setting
* Restore alphanum sorting
* Fix server-side sorting which was broken by the trackerelement rewrite
* Eliminate one extra buffer use
* Add pps option for pcapfile replay
* Add \'override\' notation
* Fix extreme aggregator to never select empty fields
* Add messages and location API to KismetConnector
* More enhancements
* Change packet too large logic to bail on size of buffer, not available
* Collapse dbm/rssi into single set of records to save memory
* Silly error not checking empty packet state
* Fix copypasta mistyping on the new double:double map
* Throw an exception on out-of-bounds packet components instead of leaking ram
* Stop using legacy _PCM reference to gps component, this is wrong and leaks memory
* Convert frequency map to DoubleMapDouble
* Fix overloaded older API
* Make pair properly
* Split CoreMap into an arbitrary-type map Add DoubleMapDouble type
* Fix gps not creating the tracked locations
* Bash autoconf into behaving with --enable-cap-sources-only
* Don\'t zero out scalar constructor value
* Fix last_signal name
* Refactor RRDs into TrackerElementVectorDoubles instead of TrackerElementVector; this saves nearly 50% RAM while retaining RRD functionality
* Abstract vector into CoreVector and add TrackedVectorDouble to reduce allocation of tracked elements for simple int vectors like those used in the RRD system
* Add ability to disable signal and packet RRDs This can reduce RAM requirements by half
* Add option to turn off packet rate RRDs
* Add shortcut function for serializing
* Switch to shared ptr for connection variables so we can handle iteration Process GET URL parameters into the connection variable cache
* Fix RRD not using aggregator
* Merge fixes
* Split up components and move them into the component tree to save on ram during compiling
* Remove const from tracked vector erase to make older compilers happy
* Mute warning on old mutex
* Use std::move to on cloning to make older compilers happier
* Add signatures for most other trackercomponent based complex elements
* Start adding signatures to tracker_component pieces
* Fix datasource driver mapping to use fixed name field
* Fix initialization on DeviceKey Fix up API usage for various null pointers on dynamic fields Initialize numerics to 0 Fix various API usage for mixed types Initial functional levels
* Bootable but incorrect json output
* Use global entrytracker
* Fix instant abort due to referential loop fetching mendatory entrytracker in http endpoints Remove entrytracker from http endpoint
* More API fixes
* Moved templated functions into header class where they had to be in the first place Split trackedcomponent off from trackedelement for ordering reasons Fix destination casting for new dynamic destinations
* Compiles, does not link
* Further significant API improvements Removal of most references to entrytracker as a shared pointer Further simplify creation of tracked fields Add safe casting mechanism with type validation on fields Further API conversion
* Update .gitignore
* Python 3 compatible print statements
* more API work, devicetracker & related now compileable
* More work on new API
* More API conversion
* More API conversion and globalreg stripping
* more migration
* Start phasing out passed globalreg More conversion towards new tradked fields
* More conversion to new API
* More global variable name collapsing More progress towards compiling with new tracked element API
* Update streamtracker
* More conversion
* Update kis_datasource for new tracked element api Add some element initialization shortcuts
* Progress towards new tracker format Bulk conversion of ENTRY_TRACKER tag to ENTRYTRACKER
* Get tracked location compileable
* Additional shortcuts to the global namespace globalreg
* Added basic math operators for CoreNumerics
* Further tweaks for new api
* Switch to enforcing a TrackerElementMap on all sub-elements
* Adapt to new tracker format
* More rewriting, change of some methods, addition of smoother registration processes for new fields
* New entrytracker and trackedelement now buildable
* Set IDs properly in new constructors
* Work converting entrytracker to new model Clean up checksum code slightly More work on new entry tracker model
* More conversion to the new model
* More work on rewrite
* Start major rewrite of internal structures to eliminate shared union and make the whole system simpler to use and more C++11-idiomed
* More reduction of shared_ptr duplication by inline make_shared
* Define some more default copy/move restrictions
* Squash some clang warnings
* Detect clang Switch to using from typedef Fix typos
* Generate default move for trackerelement and delete default copy
* Remove constexpr because very old (debian) g++ can\'t handle them properly
* More minor C++11ism optimizations
* Tweak loops to use marginally more efficient const auto iterators
* Revamp tracked elements to use const references whenever possible to minimize some memory and temporary variable thrash
* Stop using EOL-locked mutexes in destructors as the thread behavior of blocking on a destroyed mutex is undefined
* Use std::make_shared for more of the shared pointer use; this should be faster and more memory-friendly than reset/new
* Added presizing vector option to speed up initial discovery
* Tweak core mutex locking system slightly, restore use of C++11 mutexes in modern cases
* update comments to be accurate save core file
* Moved _MSG to Globalreg:: namespace
* Only split channels between sources with exactly equal channel lists
* Added stub fmt header w/ required include-only define Added integrated libraries credits to readme Added fmt{} enabled _MSG_xyz macros
* Added fmtlib from https://github.com/fmtlib/fmt
* Fix source number calculation (uuid checksum) aliasing all sources to \'0\'
* Squash some python verbosity in helper modules
* Use self-chosen name (via --source=wlan0:name=selfchosen) in if_name field of interface description block
* Setuptools req\'d
* Moved setting of commonname into lambda options on the set functions, instead of a serialization-time live check, this may help alleviate some contention during serialization
* Remove pcap download link b/c it caused too much confusion, will restore as a different method in the future
* Remove persistent storage code for now
* Properly filter \\t, \
, \\r, \\f, \\b, etc in JSON export
* Start framework for new onboarding
* Adapt to kismet remote 2018
* Add a ms-timered conditional block option Continually re-check blocked streams for safety
* more minor potential memory leaks
* Fix possible leak in channel handling in linux_cap_wifi
* Fix possible memory leak in ppi radio header
* Fix possible (but unlikely) memory leak of ringbuf ptr
* Fix possible memory leak of
*
* pointer for sub-interface list
* Clean up probe behavior
* Explicitly set memory
* Fix possible null ref
* Fix re-enabling hopping after a lock caused by sending the channel AND chanhop records in a configure report simultaneously
* Handle restarting remote capture sources more cleanly
* Docs
* Python cleanup
* Cleanup of python code
* Docs on upcoming rest api for dblog
* Fix --disable-python to the proper --disable-python-tools
* Fix broken makefile
* Patch for protobuf-c on openwrt for compiling on the host
* Explicitly check for python2 modules in autoconf
* Fix quirks in hop offset setting
* Makefile got skipped b/c of gitignore
* Revert \"Listen to DESTDIR variable in python modules for packaging (thanks to\" Doing this causes permission errors This reverts commit 3d64713ab6e361c04773c4b13b8abf73ed48d804.
* Listen to DESTDIR variable in python modules for packaging (thanks to zerochaosAATTgentoo.org and mgorny for pointing this out)
* Initialize some vector data in plugintracker ahead of time to fix bug w/ disabled plugins
* Finish transitioning rtl433 to new common json report
* Split rtl433mqtt properly
* Strip out rtl433-local versions of the protobuf code Switch rtl433 to using the new json field in the datasource report
* Add json and protobuf packinfo fields, template access
* Add fixed GPS to remote sources
* Add optional high-precision time field for GPS Note that we\'re going to transition from dedicated fields in linuxbt and rtl433
* Enable remote in pcapfile so it doesn\'t silently fail
* Add SubJson and SubBuffer to the datareport field to simplify data sources
* Remove old cygwin code which is hopefully dead
* Move some more old code out of tree
* Fix misuse of set_rtlid
* Wrap http generator in try/catch
* Fix bug in new json export processing
* Fix naming on dot11 client field
* Super basic initial elk work
* Change JSON serialization to skip empty elements instead of serializing a 0 in their place; consumers of json should now check for the presence of a key rather than the key being 0
* Remove hardcoded softfloat
* ubox git url not reachable anymore (404)
* fixing curl command for automake-perl patch
* Fixed typo in CryptToHumanReadable function where it failed to reference crypt_unknown_nonwep
* Tue Apr 17 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1523843792.20581d1f:
* Catch broken rtl8814au driver
* More handling for the variant 8814au driver
* Set IPC name
* Fix starting from defined source
* Tweak package makefile
* Rename dir again
* Update docs
* Start docs update
* Bail sooner when librtlsdr is not available
* Push errors directly at the datasource if the IPC fails
* Fix exception name
* Fix typo
* Copypasta
* Send error reports
* Fix exception states
* Return a config success on rtl433
* Dont\' initiate spindown on the python ipc side, let kismet kill us
* Quiet some spurious IPC errors on datasource IPC
* Fix segv when missing l1info
* Start reworking rtl433 parsing system to handle more device types
* More work on 433 switches
* Add templated functions
* Use new functions
* Parse switch state for rtl433
* Remove some rtl cruft
* Disable uclibc since we don\'t build w/ it anymore anyhow
* Turns out protobuf 3 on openwrt doesn\'t like something named Shutdown at all
* Update for new Shutdown message name
* Fix OUI glomming
* Split tetra build off again because of openwrt differences in modern versions Fixed host deps for openwrt
* Check for python-setuptools during autoconf
* Mon Apr 02 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1522682062.5cd1df22:
* Update API
* Mon Apr 02 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1522630002.6c838f3d:
* Add open command Initialize KDS commands
* Start implementing external interface Fix makefile deps
* Derive from object
* Split KDS from KEI in python
* re-merge into one file
* Fix pub/private functions
* Start refactoring rtl433 source to a proper datasource Start rewriting rtl433 capture tool to new datasource code
* Change API to callback functions instead of searching for child functions
* Start migrating to class implementing datasource+callbacks
* Set daemon mode on the helper thread
* Refactor into module+tool, start adding protobuf python code
* Add python2 binary / --with-python2=... configure argument
* Add simple makefiles to autobuild python modules
* Generate modules config
* Update make install
* Link error report to graceful shutdown
* More work towards proper datasource for the rtl433 sdr
* Graceful spindown
* Regen sdrrtl433
* Define timestamps
* Clobber time on remote sources properly
* Clobber timestamp appropriately
* Update for timestamp
* Initial working rtl433 python standard datasource
* Fix bugs in sending error reports
* Update datasource code for rtl433 to take python from the new proper datasource api
* Update rtl433 readme
* Add --disable-python to configure Install python modules automatically during make install/make suidinstall
* Updated readme
* Tweak deps on protobuf
* Fix source parsing
* Tweak rtl deps
* Rewrite python modules install Remove sqlite3 dep from kismetlog
* Rename tetra package to openwrt Update docs for tetra and osx Disable python in openwrt builds
* Add default type string to making a new device record Set default types for all of wifi, bluetooth, etc
* Tweak deps on protobuf again, hopefully solving some openwrt build problems with remote-only
* more tweaks
* Fix removing rtl433 Tweak deps
* Mark more files as precious
* More files as precious
* Fix monitor_bin in rtl433 makefile
* Update readme
* Rewrite how callbacks are called
* rewrite how callbacks are called
* Tweak static methods Use send/recv in remote sockets Fix remote sockets
* Enable remote source for rtl433
* Add protobuf to remote makefile
* Set protoc-c binary
* Updated date
* Tweak packaging more
* (broken) mqtt support
* Fix MQTT
* Split rtlsdr mqtt into its own type def so that we can display it more accurately
* Unlock before calling callbacks
* Clean up some threading
* More threading tweaks
* More thread tweaks
* Roll back some thread changes
* Revamp how ipc tracks instances
* more work to prevent deadlocks
* Minor other DST threading tweaks
* Add pkg-config
* Fix channel hopping getting permanently cleared
* Ignore pyc
* Remove old, incorrect comments about datasource
* Remove stray mqtt import
* Remove stray requests import
* Initial work on a freaklabs zigbee datasource
* Handle sending config responsees
* Implement basic channel hopping
* Simplify hopping control
* Add basic data report sending
* Initial capture support
* empty docs
* Add docs Turn on installing freaklabs zigbee
* Play w/ docker options
* Add python deps
* Generate zigbee makefile
* Nuke source on rebuild
* Tweak docker more
* Dockery stuff
* Fix old ret val being used in tcpclient
* More dockering
* nope
* Scrap docker-compose for now
* Docs
* Add site.conf override
* Fix bluetooth not working for remote cap
* Docs, helper script
* Updated docs
* Never update signal for destination devices on wifi, only add location if location is emtpy
* Always merge a datasource w/ a known type even if opening it resulted in an error; this solves defining sources and launching kismet before the sources are present
* Update docs
* Sun Mar 25 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1521862520.3a36c344:
* Fix deps in makefile for generating protobuf Migrate all server code to protobuf implementation for datasources/remote cap Split external interface into basic and http-enabled classes Make datasources a type of external interface
* Rough start of new tracked element filter grammar
* Start generating protobuf-c code
* Add total size api
* Start converting C code to protobuf
* Remove msgpack from capture framework Fix any number of bugs in new protobuf implementation Tweak protobufs Transition datasource capture framework code to new protobuf code Remove old simple datasource proto
* Build protobuf_c for linux bluetooth Handle hop_rate properly Don\'t send hop list in report if we\'re not hopping Convert linuxbt to protobuf
* Started updating osx to protobuf
* fix api call
* OSX protobuf fixes
* OSX fixes
* Properly clean captureframework.a
* Fix tetra packaging dir
* Add debug info
* Don\'t check microhttpd version in remote-only
* Booleans are hard
* Remove included msgpack Remove built-in msgpack serialization
* Docs for new datasource api
* Detect errors and handle them properly in kismet databaselog
* Docs
* Try to look for protobuf-c0 on old systems
* Tweaks
* Docs
* Rename python_tools
* Move python rest to python libs
* Started adding KismetLog module
* Update docs
* Signedness fix
* Update protobuf compile
* Add rtl433 datasource proto in prep for new datasource Major work on the KismetExternal python module to streamline its use, add documentation, and to add datasource support
* Remove junk __main__ from python
* Mon Mar 19 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1521475215.32b4450e:
* Start adding TPMS support for rtl433 Map moisture to humidity for rtl433 Add lux and uvindex mappings for rtl433 weather stations Convert rtl433 ID to string to map complex IDs
* Force python2
* Default to metric if no preference set
* Tweak the amount written to always reflect the amount peeked in tcpclient2
* Switch socket to making an IPPROTO_TCP socket
* Squelch error in strict_fstream
* Refactor how we handle errors in tcpclient Switch to recv/send from read/write, and include MSG_DONTWAIT even tho we\'re on a nonblocking socket; this might fix OSX being broken?
* Remove some copypasta
* Refactor tcpserver to use recv/send
* Commit proper buffer quantity
* Refactor cap_framework to use recv/send when in tcp mode
* Fix a bunch of potential memory leaks and other errors reported by the new clang
* Consistency tweak
* Enable GPS snapshotting into the \'snapshots\' category of the kismet log; functions as the gps position track
* Began very rough start of KML conversion for kismet log files
* Added strongest point plotting, ssid matching, and device name plotting to KML
* Start adding protobuf requirements, definitions, etc, in prep for shifting datasource away from our custom protocol and into protobuf
* docs
* Catch some errors in some files
* Start fleshing out new protobuf based communications protocol & helper class
* Remove auto-generated python file Start building protobuf network code
* Started mocking up the HTTP tunneling protocol in protobuf
* Split http protobuf into its own definition
* Shuffle http
* Start of external api
* Stub functions
* Work on http proxy
* Re-spin http protobuf to not use maps
* More tweaks to protobuf http proto
* Move protobuf check
* Update osx howto
* Recompile protobuf parsers with protoc-2.6.1
* Remove pregen protobuf files since they don\'t function cross-pb version
* Rework makefile and autoconf to dynamically compile protobuf code since 2.6 can\'t build on 3.0 runtime and 3.0 can\'t build on 2.6 runtime
* Mark protobuf generated files as precious
* Revamp, again, how protobuf is managed in dependencies and compiling
* Add arbitrary user info fields to data sources
* Shortcut atomic writes from string buffers
* Flesh out most of the HTTP proxy protocol
* Handle HTTP-related external messages
* Moved kismet datasource / external helpers to the bin dir instead of the kismet_capture_tools to make some distrbutions happier and to put remote cap into the users $PATH
* Update capture_sdr_rtl433 readme
* Start working on python external tools library
* More work on external python library
* Move to modules dir Add setup.py
* Fix protobuf name
* Remove old rtl433 code
* Start adding test plugin for external api
* Implement first POC of proxying web requests
* Fix logic errors
* Start implementing POC of http request api in python library
* Add --with-protoc and --with-protoc-c for crossbuilding with specific protoc compilers
* Fix --with-protoc/cc options
* Start tuning for protoc on openwrt; still some errors
* Handle pipes closing in python interface library
* Rewrite proxytest as a python module, with better comments
* Start adding docs for helper interface
* Updated docs for tetra/openwrt build
* Added external api auth request/auth token exchange
* Nuke old rtl433 dev code that we\'re not using
* Handle message sending over external api
* Document the external helper api more fully
* Handle shutdown packet
* Sun Mar 04 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1520118761.70ce4d5a:
* Switch to pure-connect mode instead of bind-connect in tcpclient2
* add EWOULDBLOCK test for completeness
* Add some debug
* Update docs for rtl433
* Tweak web docs Add timestamp to stream tracker
* Update database to v4; include type in DATA section Log rtl433 JSON data as a DATA log in kismet log
* More refinement of the table refresh
* Update datatables Add a debug for weird ui data ranges Change main view update method to try to fix a scrolling bug
* Hardcode pthread; we can\'t work without it and we don\'t work on a system that doesn\'t have it
* Tweak remote pkg too
* Switch to ax_pthread detection again, fix openwrt w/ a hammer
* Remove old CLIBS
* Include cmath
* include cmath header
* Revert \"Switch to newer ax_pthread test\"
* Revert \"Mess around with pthread linkages\"
* Mess around with pthread linkages
* Switch to newer ax_pthread test
* Fix make clean typo
* Use std:: namespaced isnan/isinf
* Remove C++isms from end of config.h; don\'t use std::snprintf globally
* Add proper return val in devicetracker virtual stub
* Huge refactor of all std:: namespace items into the proper namespace, remove \'using namespace std\' entirely
* Add test for std::isnan Fix \'make clean\' nuking rtl433 capture
* Moved database docs from code to md
* Fix regression in pcap streaming
* Fix setting source type in datasource list
* Add info msg with ssid
* Add device key to devices table in kismet log
* Mon Feb 26 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1519662292.77f9007f:
* Log datasources in the kismet log at intervals & on creation of a new one
* use the kis_log_packets config file option properly
* Default to \'true\' for logging devices, if the config file hasn\'t been updated
* Set rtl hardware
* Revamp some of the datasource proto for dynamic array conversion Add chipset/hardware reporting to probe/list/datasource to show chipset in linux, etc
* Remove unused errstr
* Revamp locking on streaming req completion to always unlock regardless of error code of streaming generator Switch back to scope locks in streaming buffer data pending functions Move conditional lock inside aux scope lock to protect from lock/unlock sequence problems in streaming buffer data pending
* Corrected \"return false\" to \"return False\"
* Revert \"Add warning about rtl8192cu\"
* Add warning about rtl8192cu
* Fix name in AP and Client records
* Add \'kismet.device.base.commonname\' which is synthesized from user name, device name, or mac address, for simple search/filter operations
* Call pre/post serialization during simplification to trigger active codepaths
* Fix multi-line?
* Flag passive sources as a passive state
* Make OSX happy by avoiding std::isxxx
* Handle un-indexable ht80/ht160 channels properly
* Typo
* Add autoconf test for old deb jessie microhttpd that breaks form decode
* Fix memory leak of groups buffer Check uid == 0 Check gid == file Don\'t print \'no groups\' error, jsut fall into \'not in the right group\'
* Remove stray debug
* Copypasta
* Add MQTT mode to rtl433 bin Add better error handling / runtime lifecycle
* Add debug switch to rtl433 Improve rtl433 error handling Add interface definition to rtl433 source Raise warning on rtl433 source creation Disable \'retry\' on rtl433 datasource since it means nothing
* Add kismet_cap_sdr_rtl433 to install
* Rename for consistency
* Rename binary for consistency
* Added help
* Reconfigure source during reload
* Fix temperature decode on rtl433 Add initial rtl433 capture code
* Start rewriting some of docs format
* Clean up alertracker
* Return created datasource in response to add_source
* Sun Feb 11 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1518318590.bd325efe:
* Comment out rmm ie tag parse fail to prevent spam
* Flesh out API in datasource_rtl433, move json handling, turn phyrtl433 into a packet handler
* Catch errors parsing IE
* Fix flagging dot11 packet types
* Sat Feb 10 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1518215167.85a798f5:
* Start implementing new proper rtl433 data source
* Fix over-aggressive consumption of by-uuid URIs
* Fix passive source opening
* Patch jsoncpp for old platforms
* Add std::snprintf shim for old, broken openwrt builds
* Migrate from nlohmann modern c++ json to older jsoncpp parser
* Added jsoncpp amalgamated json parser in preference to the nlohmann json.hpp parser; too many systems still in use can\'t handle the modern C++isms
* Fix \'if\' block
* Add population of empty device records for wifi destination devices only seen as tx-to
* Add has_ block to dynamicproxy tracker element
* Typo
* Protect against nan/inf double/floats in json serialization
* Rename historic fields to kismet.foo instead of kis.foo Protect divided alts from returning nan
* Add server_name, server_desc, server_location strings to configs Add server name, desc, location, and uuid to system monitor output
* Add server UUID to device record
* Process WPA handshake on source path as well as dest path
* Basic wifi client count
* Add counter fields populated during preserialization for number of wifi clients, etc
* Fix AKM cipher vector
* Thrash
* Mangle serial
* Don\'t just abort on weird data
* Fix OUI extraction Fix populating tag stream
* Fix recursion
* Fix ie221 subtype
* Kismet compileable with new parsers
* Start porting to the new parser api
* Move to dot11 parser tree
* Rename ht to ht_cap for consistency Clean up some static variables Fix various API mixups
* Name file properly
* Fix shred ptr in ie 45 Rename ops for consistency
* Add dot11 action
* Port dot11 action frame
* Add revamped parsers to makefile
* Port wpa-eap parser
* Port ie7 country code (which clued us into the kaitai problem in the first place)
* Ported ie221 wfa wpa
* Port ie221 ms wps
* Minor formatting
* Port ie221 ms wmm
* Add stream to ie
* Port 221 generic vendor
* Port ie 221 droneid
* Fix API
* Remember proper API
* Port IE192 VHT Op
* Port IE191 HT CAP
* Port IE133 Cisco CCX
* Port IE61
* Refactor to use shared ptrs
* Add IE54 mobility Add IE55 fastbss
* Port IE52
* Fix header defs
* REfactor bool to uint
* Add partial RSN
* Convert IE48
* Rewrite ie45
* Rewrite dot11 QBSS
* Minor refactor
* Start porting parsers to custom code
* tl;dr - kaitai as of 2018-01 generates bad C++ code which leaks memory both under switch() data blocks (avoidable) and during exceptions in constructors (unavoidable).
* Add ulimit config to force an abort on memory growth
* Properly handle un-study-able pcre
* Wed Jan 31 2018 mardnhAATTgmx.de- Update to version 0.0.0.git1517339966.ee6d9464:
* Disable dot11d country parsing for now because it exposes the worst in the kaitai memory leaks
* Possible workaround for memory leak in kaitai code generation under C++
* Solve some memory leaks in the use of dynamic kaitai objects
* Hack in a fix for std::to_string on old (openwrt) compilers
* Add add_channels source option for wifi, add docs on channel definitions
* Don\'t try to upgrade unopened database
* Look for group membership when running ipc binaries
* Include channel= in channels definitions of a source if the channel isn\'t in the autodetect list
* Convert channel to frequency in set_freq via nl80211
* Fix purging of devices under max_devices count
* Don\'t leak buffers in some error conditions
* Handle null ptr in device list due to timeouts
* Remove debug string
* Don\'t double-close dumpfile
* Don\'t allocate a 0-length buffer
* Unneeded, and wrong, uppercasing
* Initialize buffers
* Fix missing initializations leading to potential garbage
* Change new\'d ptr to unique ptr Fix some indenting
* Fix leak of ppi radioheader on corruption
* Fix leak by not assigning decrypted data over datapayload
* Fix copypasta
* Restore destination management packet device type classification
* Tweak the adhoc/client/etc wifi classifier
* Added TCP NMEA GPS client to use smartphone NMEA apps as GPS
* Split NMEA into a superclass in prep for adding nmea over tcp
* Fix possible /0
* Typo
* Fix a multitude of potential null pointer dereferences found by kevin bryan w/ clang
* Updated the README regarding known issues on some wifi sources
* Clean up linux wifi probe code to skip known unusable interfaces Add option to skip HT and VHT channels as a source option Detect iwlwifi and disable HT and VHT
* Add kismet.AddLoadComplete() callback for adding post-dynamic-loading CBs Add kismet.LoadDynamicIncludes() to handle the promise resolution with an automatic callback
* Handle channels= options to set custom channel lists
* Namespace fixes for std::
* Add string handling functions for quoted parameters (for channel lists) and for merging unique channels
* Update comments to reflect move to handle custom channel lists in the capture source driver
* Update code to match proper documented names and for consistency
* Update readme to reflect use in code
* Fix segv where trying to set channel before opening mac80211 when channel= provided in source definition
* Set ht40 widths properly and more consistently with how \'iw\' does it
* Set channel type in netlink commands on linux
* Set translated telemtry GPS objects as \'valid\'
* Add username to system status
* Use thread-safe pwent operations
* Pull in WFA-WPA IE tags and process them for beacon crypt
* Rework kaitai definitions for wfa
* Set model for mavic broken firmware
* Identify broken mavic drones based on corrupted droneid
* Remove blocked channels from readme for now
* git clone for openwrt changed
* building package of remote capture for OpenWrt
* Convert gcc error to warning because of ubuntu 14.04
* Rewrite how columns are processed in datatable sort
* Fix empty notes breaking UI
* Add newline-escaping/restoring UI code Add per-device notes
* Import MIT-licensed C++ JSON parser from https://github.com/nlohmann/json Remove homegrown kismet JSON parser Convert gpsd JSON, rtl433, and zwave JSON parsers to nlohmann JSON Write nlohmann JSON parser into Kismet Structured common layer
* Add arbitrary tag API
* Fix wifi freq of 0
* Remove debugging
* File file write buf not handling fwrite returns cleanly Add msg output to pcapng Activate pcapng logs
* Fix not handling return values properly & always defaulting to /dev/random instead of urandom
* Add initial pcapng file log
* Add file output buffer
* Typo
* Fix OSX HT40 channel tuning
* typo
* Use a FILE
* for ppilog and set it to clo_exec
* Fix sizing of thread stack on disabled thread code (correct even if disabled)
* Use proper logfile generators
* Use on-disk journal but clean it up properly
* Close database properly during dblog close
* Fix massive corruption bug caused by calling randombytes on a uuid twice
* Don\'t adjust thread stack size
* Switch to a pthread for the packetchain thread so we can force a sane stack size on some platforms (musl libc)
* Split up ac_find_libs into two lib checks
* Remove sysvinit comments in systemd Run w/out ncurses in systemd
* Add configure-time detect for mhd_quiesce being available Restore mhd_quiesce when possible; some microhttpds lack that, others hang during mhd_stop so it\'s a lose-lose
* Update kismet_uav.conf.yaml
* Updated docs for compiling for tetra-2.0 firmware
* Include zlib dep on tetra
* Fix use of strerror_r on uclibc, others
* Hacky workaround for uclibc configfile globbing
* Update openwrt tetra package
* Revert \"Remove libpcre in favor of glibc regex; we don\'t need the extra PCRE\"
* Add Spark and Spark RC to the uav matches Fix telemetry showing when it is empty in uav details Hide serial number when it\'s unknown in uav
* Don\'t generate client assocs for probe, assoc, etc reqs
* Add altitude to distance prefs code
* Rework droneid ksy a bit more to match other methods Support extracting serial number from otherwise flaky/broken 0x11 subcommand droneids; dji devices set memory wrong and send a corrupt frame
* Allow field specifications for device method
* Rewrite kaitai bitfield parser to match other parsers Fix telemetry display for droneid in the webui
* Fix typo
* Migrate UAV configs to YAML format which is \'compiled\' into a kismet config (prep for future fingerprint development in a similar format)
* Remove libpcre in favor of glibc regex; we don\'t need the extra PCRE stuff, and we don\'t need to link another library
* Solo is made by 3dr
* Sat Dec 30 2017 mardnhAATTgmx.de- Initial package, version 0.0.0.git1514605988.c62a9d35