dump1090-mutability (1.15~dev) UNRELEASED; urgency=medium * In development. * Validate that the username given to debconf is syntactically OK and isn't root. (github issue #24) * Don't fail on postinst if the given user exists but is outside the usual system user UID range. (github issue #24) * Fix timestamp correction when sample blocks are dropped. (github issue #43) -- Oliver Jowett Thu, 19 Feb 2015 22:39:19 +0000 dump1090-mutability (1.14) unstable; urgency=medium * Position reporting: * Use the correct maximum inter-position-message intervals for surface position messages (github issue #18). * Disable receiver-relative position decoding entirely if the receiver max range is larger than 360NM. * Implement speed checks on position updates; implausible position changes are discarded, unless they improve NUCp. Fixes most cases of "spikes" in position tracks for aircraft with poor GPS data, and helps with many cases where aircraft are transmitting garbage position data. (github issue #17) * If a position update is considered implausible, discard all the other data in the same message too, just in case - often altitude will be wrong in this case, too. * Message decoding/forwarding: * NUCp (navigational uncertainty) is extracted from position messages (github issue #16) * Improved support for DF18 ADS-R/TIS-B message formats and cases where they use non-ICAO addresses (handle the IMF bit) * Don't forward the first message for an aircraft to network clients until we have seen a second message from the same aircraft. This should handle most of the "garbage ICAO address" cases. Disabled in --net-verbatim mode. * Aircraft with only one message seen are pruned more aggressively (60 second timeout vs. 300 second timeout normally) * Don't emit SBS output for non-ICAO addresses (github issue #9) * Improved the calculation of message reception time seen in SBS output. Now message reception time should be strictly before current time. * Don't try to correct more than one error in DF11 messages as these messages have reduced CRC coverage and multiple-bit errors are ambiguous. * Fix some undecodable altitudes that would be interpreted as a valid altitude of 0ft. * Stats & JSON: * Documented the JSON formats (see README-json.md) * Fix "-infdB" in stats if no signal peaks have been recorded. * NUCp is reported in aircraft.json * Fix "unrecognised ICAO address" stat counter to count the right thing. * Gather stats on number of unique aircraft tracks seen, and number of tracks where only a single message was seen (which are probably garbage). * Fix stats timestamp output during the period shortly after restart. * Include timezone in timestamps shown with --stats * Add various stats for CPR position-decoding details. * Webmap: * Fix webmap history loading when no history is present. * Better handling of long refresh times in the webmap. * Packaging/misc: * Fix config handling of lat/lon between -0.999 and 0 (github issue #14) * Add --oversample option to help text (github issue #11) * More options now accept fractional seconds (--stats-every, --net-heartbeat, --net-ro-interval, --write-json-every) * Start dump1090 with a niceness of -5 to reduce the chance of dropped sample data, except in --net-only mode where it's not using a RTLSDR dongle (github issue #19) * Make sure background tasks have a chance to run even if the sample rx thread is blocked waiting for a USB reconnect. (github issue #20) * Include timestamps in logging of USB disconnects/reconnects (github issue #21) * Catch SIGTERM and do clean shutdown like SIGINT does. Also log about it. (github issue #22) * Log (with timestamps) on startup/shutdown so it's more obvious in logs where dump1090 restarted. * Mark all HTML/Javascript as conffiles to avoid overwriting user changes on package upgrade, as many people seem to modify these files. (github issue #15) -- Oliver Jowett Thu, 19 Feb 2015 19:41:39 +0000 dump1090-mutability (1.13) unstable; urgency=medium * Moved to a simpler versioning scheme rather than trying to put suffixes on the end of the upstream version scheme. * Statistics updates for the stats junkies amongst us. Note that many of these apply to the 2.4MHz demodulator only. * Rationalized the demodulation stats that are collected. * Collect mean signal level, peak signal values. * Collect a count of signals stronger than -3dBFS. * Collect stats on remotely-received messages. * Collect stats on CPR position decoding. * Collect more CPU stats. * Stat buckets for the last 1, 5, 15 mins are maintained. * Stats display due to --stats-every does not reset the "total" stats. * All stats are now exported via JSON at /data/stats.json. * When --oversample --phase-enhance is in use, all phase positions are evaluated and the best result is used, rather than stopping at the first error-free message. This increases CPU by about 1/4 but recovers about 5% more messages. * Error correction in --aggressive mode is now more conservative: it will only fix 2-bit errors that can't be confused with a 4-bit error. * DF11 messages can now be partially error-corrected (only for errors that do not result in a syndrome in the final 7 CRC bits) * Partial TIS-B support. Feedback sought from people who actually receive TIS-B messages! * Support for non-alphanumeric characters in callsigns. * RSSI measured for individual messages. * Simple filtering for bogus messages that make it past CRC checks: don't display or emit aircraft state until we have seen at least two messages from the same aircraft. * --net-verbatim option added; if enabled, for messages that have repairable CRC errors the unrepaired version of the message is forwarded to network clients. This lets the client decide what policy to impose on damaged messages. * Made 2.4MHz sampling the default for new installs. * Internal changes: * Restructuring of the code to be a bit more modular. * Mode S message decoder and output formatting cleanups. * 2.4MHz demodulator main loop unrolled to demodulate a byte at a time; this helps compensate a little for the --phase-enhance change above. * CRC calculation switched to a faster byte-at-a-time table lookup. * ICAO address filter rewritten to be a bit smarter, and to support Comm-B overlay control in the future. * Non-ICAO addresses (e.g. from TIS-B and ModeA/C) are flagged internally so they cannot be confused with a real ICAO address. * More changes to try to work around libusb crashes on exit. * Webmap changes: * Show recent RSSI for the selected aircraft. * Non-ICAO addresses are shown italicized. * Fix webmap stable sort of rows with no sortable values (would cause "jumping" rows at the end of the table) * Doubleclick on an aircraft marker or table row to follow that aircraft, recentering the map as it moves. Or click on the selected aircraft's callsign or the arrow next to it. * config.js cleanups. Page title and site name can be specified here now. * Added vertical rate indicators to altitudes on the webmap. * Linked from the aircraft ICAO to airframes.org. (Other site recommendations welcome!) * Infoblock layout changes. * License for my modifications changed to GPL2+. Previous binary releases were effectively already GPL2+ due to linking with librtlsdr. -- Oliver Jowett Sun, 25 Jan 2015 16:45:00 +0000 dump1090-mutability (1.10.3010.14mu-12) unstable; urgency=medium * Add generation of history JSON. * Load history when the webmap is initialized. * Display stats on exit in --net-only mode. * Explicitly pick the closest tuner gain to the one requested. * Include a status message in HTTP response headers. * Fix sort-by-message-count in the webmap. * Always sort missing values at the end of the table, even in a descending sort. * Avoid a (rare) divide-by-zero crash in 2MHz --phase-enhance. * Try to work around a libusb crash on exit after SIGINT. -- Oliver Jowett Sat, 17 Jan 2015 21:03:55 +0000 dump1090-mutability (1.10.3010.14mu-11) unstable; urgency=medium * Fix UTC clock display in non-UTC timezones. * Layout tweaks in the info panel. * Change plane color if we have not seen a recent position update (previously, any type of message was enough). * Fix thinko in skipping over decoded message samples in the 2.4MHz demodulator. * Fix a couple of problems found by valgrind. * Fix several problems with surface position decoding and relative position decoding. * Fix hang on exit after SIGINT. * Add --max-range parameter. Use it for relative position limits, and to discard bad position results. * Display message rate (30 sec average) on the webmap. * Restart lighttpd on package upgrade if we changed its config files. * Require that the JSON refresh interval is at least 1, as it is used to control the webmap refresh rate even if JSON files are not being written. -- Oliver Jowett Wed, 14 Jan 2015 01:17:51 +0000 dump1090-mutability (1.10.3010.14mu-10) unstable; urgency=medium * Many changes to aircraft.json and the webserver code: * Restructure to include the current timestamp, and move aircraft data down a level. If you're processing the JSON yourself, you will need to adjust your code to handle this. * Only include aircraft data when we are sure it's valid. * Include the time of the last position report in. * If the aircraft is on the ground, report altitude as "ground". * Strip trailing query strings from URLs before matching, to allow AJAX cache-busting query strings to work. * Disable the lighttpd stat cache in the provided config fragment, to avoid problems with partial content being served if lighttpd caches an old size for the JSON data. This would manifest as "bad content length" or JSON parse errors in the webmap. * Many changes to the webmap javascript: * Internal restructuring of how the plane table and map markers are maintained. Should be faster and less memory-hungry. * Sorting the plane table now sorts missing data high (end of the table for an ascending sort). * On-the-ground aircraft display an altitude of "ground". * Dashed tracks are used when there are lots of missing position reports (more than 5 seconds between positions). * Green tracks are used for on-the-ground aircraft. * Limit history storage requirements by only retaining at most one position per 5 seconds. * Removal of old aircraft should be much more reliable now, even across long pauses (e.g. browser hibernation). * Change the clocks to be two UTC clocks: one from local browser time, one reflecting the receiver timestamp in the last update. * Add "total aircraft" and "total aircraft with positions" counters to the infobox when no aircraft is selected. * Add a "history positions" counter to the infobox when no aircraft is selected; this shows the number of points making up the historical track lines. * When an aircraft is selected, show both metric and imperial units in the infobox (metric setting controls which is shown first). * When an aircraft is selected, show the time since last message and time since last position message in the infobox. * If aircraft.json can't be retrieved, display a warning. * If aircraft.json can be retrieved but doesn't appear to be being updated (e.g. external webserver, and dump1090 has stopped), display a warning. * config.js: Change the default map zoom setting to be more zoomed in. * config.js: Add some comments. * config.js: Add an option to hide the clocks. * Remove extension.js and options.js, as they aren't really maintained. * Enable strict mode for script.js / planeObject.js. * Fix initial map centering to correctly use the receiver location. -- Oliver Jowett Wed, 07 Jan 2015 23:54:10 +0000 dump1090-mutability (1.10.3010.14mu-9) unstable; urgency=medium * Fix warnings. Add -Werror so they break the build in future. * Remove dead tracking code related to the removed PlanePlotter feed. -- Oliver Jowett Sun, 04 Jan 2015 20:23:26 +0000 dump1090-mutability (1.10.3010.14mu-8) unstable; urgency=medium * Version-only rebuild to reflect the upstream dump1090 version we have merged from. Retaining the trailing build number to avoid confusion (as far as that's possible given this versioning system!) -- Oliver Jowett Sun, 04 Jan 2015 12:32:30 +0000 dump1090-mutability (1.08.2302.14+1mu-7) unstable; urgency=medium * Notice if we lose the RTLSDR device, and try to reconnect when that happens. This lets you unplug/replug the RTLSDR without dump1090 exiting, and also lets dump1090 tolerate occasional USB glitches that make the dongle reset itself. * Fix some of the more glaring pthread bugs. * Center the webmap on the receiver location. * Mark config.js as a conffile, so user changes won't get overwritten. -- Oliver Jowett Fri, 02 Jan 2015 22:53:55 +0000 dump1090-mutability (1.08.2302.14+1mu-6) unstable; urgency=medium * Add support for LOG_DECODED_MESSAGES option to log all messages (disables --quiet). -- Oliver Jowett Tue, 30 Dec 2014 17:09:36 +0000 dump1090-mutability (1.08.2302.14+1mu-5) unstable; urgency=medium * Add Provides: fatsv-data-source -- Oliver Jowett Tue, 30 Dec 2014 00:01:51 +0000 dump1090-mutability (1.08.2302.14+1mu-4) unstable; urgency=medium * Remove half-implemented --no-decode option. * Remove Windows support files that aren't needed for the Debian package. * Remove some legacy support scripts not needed by the Debian package. * Add support for controlling the accuracy of the receiver location written in the JSON metadata used by the webmap. * Oversampling is now less scary! * Warn if --modeac is used together with --oversample. * Link to github from the webmap. -- Oliver Jowett Sat, 27 Dec 2014 22:24:39 +0000 dump1090-mutability (1.08.2302.14+1mu-3) unstable; urgency=medium * Ask about json interval before json dir and skip the dir question if the interval is disabled, it makes more sense that way. * Be much more cautious about missing config settings so we don't explode so badly if something is omitted. * Use the package version as the version number compiled into the binary. * Add data/receiver.json (generated once) and support for it in script.js. * Internal webserver rearrangement to support multiple json files. * Dongle "device index" actually accepts serial numbers too, document that. * Add input validation to most of the config questions. -- Oliver Jowett Wed, 10 Dec 2014 21:58:12 +0000 dump1090-mutability (1.08.2302.14+1mu-2) unstable; urgency=medium * Fix a memory leak from use of realpath() in HTTP request processing. * Add direct support for FATSV-format output, adapted from the FlightAware fork. * Big rearrangement of configuration to add debconf support. * Update Build-Depends based on glitches found building under pbuilder. -- Oliver Jowett Wed, 10 Dec 2014 01:06:56 +0000 dump1090-mutability (1.08.2302.14+1mu-1) unstable; urgency=medium * Initial release. -- Oliver Jowett Sat, 06 Dec 2014 23:07:02 +0000