[aprssig] "no sym yet" icon + compressed packets? - FIXED on aprs.fi

[Heikki Hannikainen]

On Sat, 15 Feb 2014, Steve Dimse wrote:

> this for a while, but it is like the bug I have with base91 weather.
> Every time I start to work on it I get a gigantic headache, so I have a
> beer. Then I stop caring about symbols and base-91 ;-)

I greatly appreciate this approach, but I'd already had quite a few pints 
of good ales in London pubs this week, so I ended up adjusting the parser 
code instead while waiting for the weather-delayed flight at Heathrow :)

On Sat, 15 Feb 2014, Paul Bramscher wrote:

> I worked the ISS again this evening, and when K0GDI worked me it sure
> appears that "1c" is missing in his packets for some reason (evidently
> from the hex view).  I guess that breaks the code for icons, if not
> more.  I guess we just won't get icons when that software/version reports.

Yes, you do, since yesterday evening aprs.fi gets the correct symbol 
table identifier from packets which have had the non-printable 0x1C byte 
removed:

http://aprs.fi/info/KD0KZE

Earlier I wrote:
> The aprs.fi code that translates that to text ("House (HF antenna)") 
> seems to be a bit buggy, it probably should say something like "invalid symbol
> table or overlay code".

I fixed that bit, the hover-on text description of a symbol now says 
"Unknown symbol" if the store symbol table identifier does not make sense.

> The APRS packet code at this end also needs a bit of improvement in the 
> handling of mangled mic-e packets, I think it should be able to get the
> symbol table right, since it is present in the remaining packet data.

Fixed. It was collecting the symbol table identifier before putting in a 
dummy byte in place of the 0x1F (or other non-printable char) eaten by 
the broken K0GDI-6 igate.

The eaten binary character in this case contains course and speed 
information. I also adjusted the code to not try to decode course or speed 
if these bytes are missing, so now those keys are not shown for the 
mangled packets:

http://aprs.fi/?c=raw&call=KD0KZE&limit=50&view=decoded

> The K0GDI-6 igate should also be fixed to not eat bytes out from mic-e 
> packets.

I'll leave it to the locals to kindly inquire K0GDI-6's owner if he could 
reconfigure his igate to run the TNC in KISS mode and use a better igate 
software, such as Lynn's APRSIS32.

   - Hessu