On 23-11-14 16:25, Dirk Hohndel wrote:
On Sun, Nov 23, 2014 at 11:39:55AM +0100, Henrik Brautaset Aronsen wrote:
I did a new fresh import from my Petrel on the latest master and saved it. This time the saved petrel2.xml [1] does not contain the empty <cylinder /> statements. But id does contain cylinders with something in them, even on gauge dives:
<cylinder o2='52.0%' /> <cylinder o2='18.0%' he='45.0%' />
I haven't used my Petrel in gauge mode (that's one heck of an expensive gauge...). It would make sense that it doesn't support any gases in that mode, but libdivecomputer doesn't know that the Petrel was in gauge mode so it tells us about those gases as they are present in the dive header. I am making a lot of assumptions here - but since I know a little bit about the Petrel data format I think these are reasonable assumptions.
Jef, can you confirm this analysis?
Yes, your analysis is correct. I just checked the documentation, and the gauge mode isn't mentioned anywhere. So either it's just missing in the documentation, or it's not recorded.
The Petrel (and Predator too) is also one of the devices where the dive mode is stored in the sample data, and not the dive header. I assume that's done because you can switch between OC/CC during the dive. But that makes it also rather difficult to tell whether a dive is OC or CC (e.g. what should we do if there are mixed samples?). And thus what libdivecomputer does is just return all gas mixes.
I tried to do a data dump, but both subsurface and the libdivecomputer dump app doesn't cooperate:
IIRC, the Petrel doesn't support dumps.
Result: Unsupported operation
The Petrel protocol doesn't support memory dumps. The Petrel also supports the older Predator protocol. That one supports memory dumps, but is only capable of downloading the last 128K of memory.
For protocols without support for memory dumps, I usually dump the individual dives, using this patch for the universal app:
http://www.libdivecomputer.org/tmp/universal_dump_dives.patch
Jef