[aprssig] APRS Digipeter Design Questions

[Tapio Sokura]

Quoting AE5PL Lists <HamLists at ametx.com>:

> Technically not a dupe (hence your quotes).  The packet without the
> trailing space is a mangled packet (the I-Field has been modified
> either by a TNC or by software).  javAPRSSrvr will filter out the
> mangled packet IF it is received by the server _after_ the packet
> which has the trailing space.

What makes the packet with the trailing space(s) the mangled one and not
the other way around?

Passing the packet that comes first and filtering all subsequent
duplicates sounds more logical to me, irrespective of which packet
has the extra space(s).

> The only characters precluded from being in an APRS packet are cr and
> lf.  This is because the non-AX.25 networks, such as APRS-IS, use

To me it looks like all bytes with a decimal value of under 28 or over
127 are also excluded from appearing anywhere in APRS packets. But many
programs seem to allow the use 8-bit characters. Some don't accept them
and the result is that some filter them, some replace them with spaces
and some just pass them along. The same with bytes under 27 or 32,
depending on software (and the TNC). Also some APRS packets seem to
contain a trailing CR in AX.25 level, some don't. What a mess.

> It is important to note at this point what a digipeater is.  A
> digipeater is a AX.25 level (link level) digital repeater.  As such,
> it should do no interpretation of what is in the I-Field.

The digipeater could examine the packet for a possible third party
header so it can detect duplicates that have been gated to inet
and then gated back to RF by another igate. But I guess this
processing could be a bit heavy for a TNC-based digipeater.
Detecting these duplicates should be the work of the igates anyway,
but there seems to be many programs in use with questionable
igate functionality.

  Tapio