- Additional documentation location to slow down growth of main repository. [https://github.com/wb2osz/direwolf-doc](https://github.com/wb2osz/direwolf-doc)
- New ICHANNEL configuration option to map a KISS client application channel to APRS-IS. Packets from APRS-IS will be presented to client applications as the specified channel. Packets sent, by client applications, to that channel will go to APRS-IS rather than a radio channel. Details in ***Internal-Packet-Routing.pdf***.
- New variable speed option for gen_packets. For example, "-v 5,0.1" would generate packets from 5% too slow to 5% too fast with increments of 0.1. Some implementations might have imprecise timing. Use this to test how well TNCs tolerate sloppy timing.
- Improved Layer 2 Protocol [(IL2P)](https://en.wikipedia.org/wiki/FX.25_Forward_Error_Correction). Use "-I 1" on command line to enable transmit for first channel. Compatible with Nino TNC for 1200 and 9600 bps.
- Dire Wolf now advertises itself using DNS Service Discovery. This allows suitable APRS / Packet Radio applications to find a network KISS TNC without knowing the IP address or TCP port. Thanks to Hessu for providing this. Currently available only for Linux and Mac OSX. [Read all about it here.](https://github.com/hessu/aprs-specs/blob/master/TCP-KISS-DNS-SD.md)
- The transmit calibration tone (-x) command line option now accepts a radio channel number and/or a single letter mode: a = alternate tones, m = mark tone, s = space tone, p = PTT only no sound.
- The BEACON configuration now recognizes the SOURCE= option. This replaces the AX.25 source address rather than using the MYCALL value for the channel. This is useful for sending more than 5 analog telemetry channels. Use two, or more, source addresses with up to 5 analog channels each.
- For more flexibility, the FX.25 transmit property can now be set individually by channel, rather than having a global setting for all channels. The -X on the command line applies only to channel 0. For other channels you need to add a new line to the configuration file. You can specify a specific number of parity bytes (16, 32, 64) or 1 to choose automatically based on packet size.
- The t/m packet filter incorrectly included bulletins. It now allows only "messages" to specific stations. Use of t/m is discouraged. i/180 is the preferred filter for messages to users recently heard locally.
- Packet filtering now skips over any third party header before classifying packet types.
- Rather than trying to keep a bunch of different platform specific Makefiles in sync, "cmake" is now used for greater portability and easier maintenance. This was contributed by Davide Gerhard.
- "-X" option enables FX.25 transmission. FX.25 reception is always enabled so you don't need to do anything special. "What is FX.25?" you might ask. It is forward error correction (FEC) added in a way that is completely compatible with an ordinary AX.25 frame. See new document ***AX25\_plus\_FEC\_equals\_FX25.pdf*** for details.
- Receive AIS location data from ships. Enable by using "-B AIS" command line option or "MODEM AIS" in the configuration file. AIS NMEA sentences are encapsulated in APRS user-defined data with a "{DA" prefix. This uses 9600 bps so you need to use wide band audio, not what comes out of the speaker. There is also a "-A" option to generate APRS Object Reports.
- Receive Emergency Alert System (EAS) Specific Area Message Encoding (SAME). Enable by using "-B EAS" command line option or "MODEM EAS" in the configuration file. EAS SAME messages are encapsulated in APRS user-defined data with a "{DE" prefix. This uses low speed AFSK so speaker output is fine.
- "-t" option now accepts more values to accommodate inconsistent handling of text color control codes by different terminal emulators. The default, 1, should work with most modern terminal types. If the colors are not right, try "-t 9" to see the result of the different choices and pick the best one. If none of them look right, file a bug report and specify: operating system version (e.g. Raspbian Buster), terminal emulator type and version (e.g. LXTerminal 0.3.2). Include a screen capture.
- More efficient error recovery for AX.25 connected mode. Better generation and processing of REJ and SREJ to reduce unnecessary duplicate "**I**" frames.
- New configuration option, "**V20**", for listing stations known to not understand AX.25 v2.2. This will speed up connection by going right to SABM and not trying SABME first and failing.
- New "**NOXID**" configuration file option to avoid sending XID command to listed station(s). If other end is a partial v2.2 implementation, which recognizes SABME, but not XID, we would waste a lot of time resending XID many times before giving up. This is less drastic than the "**V20**" option which doesn't even attempt to use v2.2 with listed station(s).
- Strange failures when trying to use multiple KISS client applications over TCP. Only Linux was affected.
- Under certain conditions, outgoing connected mode data would get stuck in a queue and not be transmitted. This could happen if client application sends a burst of data larger than the "window" size (MAXFRAME or EMAXFRAME option).
- New client side packet filter to select "messages" only to stations that have been heard nearby recently. This is now the default if no IS to RF filter is specified.
- The top speed of 9600 bps has been increased to 38400. You will need a sound card capable of 96k or 192k samples per second for the higher rates. Radios must also have adequate bandwidth. See ***Going-beyond-9600-baud.pdf*** in the doc directory for more details.
- Sometimes kissattach would have an issue with the Dire Wolf pseudo terminal. This showed up most often on Raspbian but sometimes occurred with other versions of Linux.
*kissattach: Error setting line discipline: TIOCSETD: Device or resource busy
Are you sure you have enabled MKISS support in the kernel
or, if you made it a module, that the module is loaded?*
- Sometimes writes to a pseudo terminal would block causing the received
frame processing thread to hang. The first thing you will notice is that
received frames are not being printed. After a while this message will appear:
*Received frame queue is out of control. Length=... Reader thread is probably
frozen. This can be caused by using a pseudo terminal (direwolf -p) where
another application is not reading the frames from the other side.*
- The Windows version 1.3 would crash when starting to transmit on Windows XP. There have also been some other reports of erratic behavior on Windows. The crashing problem was fixed in in the 1.3.1 patch release. Linux version was not affected.
- IGate did not retain nul characters in the information part of a packet. This should never happen with a valid APRS packet but there are a couple cases where it has. If we encounter these malformed packets, pass them along as-is, rather than truncating.
- Don't digipeat packets when the source is my call.
- Command line option "-a n" to print audio device statistics each n seconds. Previously this was always each 100 seconds on Linux and not available on Windows.
- In Mac OSX version: Assertion failed: (adev[a].inbuf_size_in_bytes >= 100 && adev[a].inbuf_size_in_bytes <= 32768), function audio_get, file audio_portaudio.c, line 917.
