[aprssig] Error checking within APRS packets

[Heikki Hannikainen]

On Mon, 20 Jun 2011, Keith VE7GDH wrote:

> Guido IZ6RDB wrote...
>> So if that is happening only over APRS-IS, it should be a problem of the 
>> specific APRS IGate software and environment (either a misconfiguration or 
>> a bug).
>
> I can't think of any software that has a setting that would allow a "mis
> configuration" like this that would introduce errors.
>
> We can speculate, but some specific testing should pin down the caused.

I believe it is happening somewhere on the APRS-IS, or on the igates. A 
lot of concatenated packets come in where the two packets were originated 
on different continents.

I noted this on aprs.fi, as the appended packet appeared in the 
comment string of another station, like this:

http://aprs.fi/info/?call=VE6SA-6
http://aprs.fi/info/?call=VK4KHQ
http://aprs.fi/info/?call=F6CIA

It seems to happen quite uniformly around the world. Often the 
appended packet comes from the same igate, but sometimes not.

I suspected it could be a problem in my parser's input code, so I wrote 
another independent very simple program to just log in to the full feed 
and log it all to disk (with timestamps in front of each line). The same 
corrupted packets can be found there, so I'm pretty confident they're 
coming from the IS instead of my code.

The odd thing is that it's pretty consistent, it's often the same pair of 
stations, like this:

hessu at soul1:/data/aprsis-archive/2011$> egrep -a ',qA.*,qA' aprsis-20110416.log |grep N5FAZ|less
1302914099 N5FAZ>APOT21,ELPASO*,WIDE3-2,qAR,N5FAZ-2:!3HB9TVW>APU25N,TCPIP*,qAC,HB9TVW:=4732.31N/00736.05EIGabriele 
Basel {UIV32N}
1302928392 N5FAZ>APOT21,ELPASO*,WIDE3-2,qAR,N5FAZ-2:!315HB9TVW>APU25N,TCPIP*,qAC,HB9TVW:=4732.31N/00736.05EIGabriele 
Basel {UIV32N}
hessu at soul1:/data/aprsis-archive/2011$> egrep -a ',qA.*,qA' aprsis-20110415.log |grep N5FAZ|less
1302905392 
N5FAZ>APOT21,ELPASO*,WIDE3-2,qAR,N5FAZ-2:!3146.86N/10628.53W>169/00HB9TVW>APU25N,TCPIP*,qAC,HB9TVW:=4732.31N/00736.05EIGabriele 
Basel {UIV32N}

Most often it involves aprsd as the igate of both of the packets (probably 
due to it's use of C strings...), but not always (as above). I'll try to 
find time to check the aprsd code if it could be fixed easily. But there 
could be something else going on, too.

I'm attaching links to logs for three days, grep'ed for ,qA*,qA. You might 
be able to spot more patterns out of it if you have time to look.

http://he.fi/misc/qAqA-20110414.txt
http://he.fi/misc/qAqA-20110415.txt
http://he.fi/misc/qAqA-20110416.txt

   - Hessu