[aprssig] APRS IS weirdness

[Steve Dimse]

OK, all you APRS Internet System detectives out there, here's something to track
down...

I just got a message from a user in Germany, who noted his call appears in the
findU error page. He did a complete investigation of the problem himself
(refreshing) and saw the problem is the alternate symbol set backslash was
missing from between the lat and lon. Fair enough, but he is sending it
correctly, and in fact when you look at him through the raw.cgi, packets appear
both ways a second or two apart. When I looked at the error page to check this
out myself

http://www.findu.com/cgi-bin/errors.cgi

my first surprise was how full this page is, and with mostly French and German
stations displaying the same missing backslash! I can see why a backslash could
cause problems, as it is a commonly used escape character, for example in perl
if you want to print a newline character it is \n, if you want to print a
backslash you must place two together, \\. Still, it has never been a problem
before. First guess, an IGate is eating the backslash.

The examples he gave me showed three different IGate callsigns following the q
construct, so a bad IGate was easily ruled out. Next I thought this pointed to a
digi error somewhere, but further investigation shows the same error occuring in
direct internet connected packets, for example look here:

http://www.findu.com/cgi-bin/raw.cgi?call=DB9AZ-10

This isn't a findU problem, I can see the same bogus packets on aprsworld, nd
grepping the live internet feed yeilded more examples.

My present guess is somewhere a hub is eating the backslashes, probably a third
tier hub in Europe, explaining why this is a central European problem. The
trouble with this explanation is the hub must also be injecting a duplicate
unaltered packet. 

I just can't see how this is happening. Maybe because I just woke up and am
getting ready to work my 4th night shift in a row I'm missing something obvious.
I'd love to hear a good explanation of this, it is one of the stranger things
I've seen. It would also be nice to track down the offending hub, assuming that
is indeed the source of the problem.

Steve K4HG