Suunto download missing alerts?

Miika Turkia miika.turkia at
Tue Sep 20 21:50:59 PDT 2016

On Wed, Sep 21, 2016 at 7:11 AM, Linus Torvalds
<torvalds at> wrote:
> On Tue, Sep 20, 2016 at 8:39 PM, Miika Turkia <miika.turkia at> wrote:
>> we have a report on user forum that when downloading dives from Suunto
>> D6i, some alerts are missing. And when looking at my own logs from
>> Vyper Air, I cannot find any airtime alarms since 2012. Also, now that
>> I compare the users logs, an import from DM5 gives tank pressure alarm
>> but direct import with Subsurface gives safety stop (mandatory) for
>> the same alarm. Does anyone have a hunch on where to look?
> So this would almost certainly be a libdivecomputer thing.
> The event handling is around line 600+ in src/suunto_d9_parser.c in
> libdivecomputer.  That says
>                                         case 0x0D: // Air Time Warning
>                                                 sample.event.type =
>                                                 break;
> and if you know you should have an airtime warning, you might want to
> add a debug output thing there.
> We then translate those things inro something human-readable in
> subsurface in core/libdivecomputer.c: handle_event().
> And it looks like that translation is fine, although maybe we should
> make it be based on the symbolic numbers, rather than hardcoded. But
> SAMPLE_EVENT_AIRTIME is (21), and events[21] is "airtime", so it
> _looks_ right.
> What is the number? You should see it in the xml file, it *should*
> look soemthing like
>     <event time='41:10 min' type='21' name='airtime' />
> but presumably it says "type='13' name='safety stop (mandatory)'" instead.

My own log seems to actually have aritime properly but for some reason
this event is not shown on the graph:

  <event time='78:54 min' type='21' flags='1' name='airtime' />
  <event time='78:57 min' type='21' flags='2' name='airtime' />

The guy who reported the issue has a different case, and it might be
that the newer models of Suunto have changed something. Looking into
the DM5 database, it seems that the event number there is 37 (0x25),
that is already parsed properly on DM5 DB import. I'll ask for a
memory dump of the D6i so we get the facts and don't need to wonder if
DM5 did something in-between.


More information about the subsurface mailing list