Orphan dives with Shearwater Perdix AI

[Andrew Trevor-Jones]

Subsurface 4.9.4 on Macbook Pro 10.15.4 with Shearwater Perdix AI running V71

I load the dives from multiple computers (including Garmin Descent Mk1 and Shearwater Perdix AI).  I usually load the Garmin first.

Somehow I missed some dives from the Perdix and have holes in my dive log.  I have all the dives from the Garmin.

Nothing I do lets me get the missing dives.

For example, I had the last three dives and if I do a download, it stops after those three dives and says there are no more dives. I tried "Force download of all dives" but it still fails after the third dive.

I tried deleting all the Perdix AI data back to where I no longer have holes.  It still fails after the third dive.  The last error I got was "gasmix -1 for tank 0 doesn't match"

Does Subsurface somewhere keep track of what dives have been loaded and can I clear that?

[Andrew Trevor-Jones]

I should add, retrying takes FOREVER because Subsurface loads each and every dive BEFORE the selection window.  Why does it do that?  Shearwater Desktop downloads the list of dives and then only downloads the data for the selected dives.

[Jef Driesen]

On 24/05/2020 04:34, Andrew Trevor-Jones wrote:
> Subsurface 4.9.4 on Macbook Pro 10.15.4 with Shearwater Perdix AI running V71
>
> I load the dives from multiple computers (including Garmin Descent Mk1 and
> Shearwater Perdix AI).  I usually load the Garmin first.
>
> Somehow I missed some dives from the Perdix and have holes in my dive log.  I
> have all the dives from the Garmin.
>
> Nothing I do lets me get the missing dives.
>
> For example, I had the last three dives and if I do a download, it stops after
> those three dives and says there are no more dives. I tried "Force download of
> all dives" but it still fails after the third dive.
>
> I tried deleting all the Perdix AI data back to where I no longer have holes.
> It still fails after the third dive.  The last error I got was "gasmix -1 for
> tank 0 doesn't match"

I suspect the dives are being downloaded, but are discarded due to the above
parsing error. Can you download again with the "Save libdivecomputer logfile"
checkbox enabled, and send me that log file so that I can check this?

> Does Subsurface somewhere keep track of what dives have been loaded and can I
> clear that?

It keeps a fingerprint of the most recent download dive to be able to download
only the new dives. I don't remember where it is kept, but it is ignored when
you select "Force download of all dives".

Jef

[Andrew Trevor-Jones]

It is still dying after a few dives.

I ended up deleting all dives (from all computers) back to the last hole in my Perdix dives - 48 dives in total.  This means there should be 48 dives on the Perdix to be downloaded.

It failed after the 4th dive with "gasmix -1 for tank 1 doesn't match"

[Andrew Trevor-Jones]

Trying again, but it is a good hour or so of downloading before it fails...

[Andrew Trevor-Jones]

Failed on the third dive again.

On Sunday, 24 May 2020 15:16:55 UTC+10, Jef Driesen wrote:

I suspect the dives are being downloaded, but are discarded due to the above
parsing error. Can you download again with the "Save libdivecomputer logfile"
checkbox enabled, and send me that log file so that I can check this?

 

[Linus Torvalds]

On Sat, May 23, 2020 at 10:16 PM Jef Driesen <j...@libdivecomputer.org> wrote:
>> >
> > I tried deleting all the Perdix AI data back to where I no longer have holes.
> > It still fails after the third dive. The last error I got was "gasmix -1 for
> > tank 0 doesn't match"
>
> I suspect the dives are being downloaded, but are discarded due to the above
> parsing error. Can you download again with the "Save libdivecomputer logfile"
> checkbox enabled, and send me that log file so that I can check this?

Note that the subsurface warning about "gasmix doesn't match" is not
fatal. subsurface will warn about it and then ignore it.

So no, that's not the reason for any parse errors, and it's just that
subsurface doesn't believe in the confusing "index by gasmix" model
that libdivecomputer has, and fundamentally does not work (because you
may have the same gasmix in multiple tanks, yet track them
separately).

So that gasmix warning will _not_ then cause that "dive data import
error" situation.

There's something else that libdivecomputer reacts badly to and aborts
the parsing.

Linus

[Jef Driesen]

On 24/05/2020 10:04, Andrew Trevor-Jones wrote:
> Failed on the third dive again.
>

> INFO: Read: size=0, data=
> ERROR: Failed to receive the packet. [in ../../src/shearwater_common.c:239 (shearwater_common_slip_read)]
> ERROR: Failed to receive the response packet. [in ../../src/shearwater_common.c:379 (shearwater_common_transfer)]
> ERROR: Failed to download the dive. [in ../../src/shearwater_petrel.c:380 (shearwater_petrel_device_foreach)]

This is a read timeout during the BLE communication. The other logfile you send
me shows the exact same problem, but just much earlier:

> [667.283179] INFO: Read: size=77, data=030001FF9300766480E068DFFFF4E4050100DBDC605A5824BE6319BBDAEBDC8F2CAE0B399048A2502DA4E78F01DBDCD1BFFFE9C80A020180DBDCB4B0497CC63374B5D7B91E59441673209144A0
> [667.461461] INFO: Read: size=77, data=03015B49CF1E0381A37FFFD3801404030181696092E98C66E96CAF723CB2782CE641228940B6938E3C070346FFFFA7002808060302D2C125D318CDDED95EE479649059CC824512816D271C780E
> [670.465341] INFO: Read: size=0, data=
> [670.465344] ERROR: Failed to receive the packet. [in /Users/hohndel/src/subsurface/libdivecomputer/src/shearwater_common.c:239 (shearwater_common_slip_read)]
> [670.465346] ERROR: Failed to receive the response packet. [in /Users/hohndel/src/subsurface/libdivecomputer/src/shearwater_common.c:379 (shearwater_common_transfer)]
> [670.465347] ERROR: Failed to download the dive. [in /Users/hohndel/src/subsurface/libdivecomputer/src/shearwater_petrel.c:380 (shearwater_petrel_device_foreach)]

Can you provide the subsurface log file as well? Since the BLE communication is
done by subsurface, maybe we can find a clue there?

Jef

[Andrew Trevor-Jones]

On Monday, 25 May 2020 19:21:31 UTC+10, Jef Driesen wrote:

Can you provide the subsurface log file as well? Since the BLE communication is
done by subsurface, maybe we can find a clue there?

Where do I find that and/or how do I get Subsurface to create it?  I've looked in the manual and preferences and can't find any reference to it. 

[Andrew Trevor-Jones]

I reran the download attempt with Subsurface creating a log file as well.

This time it didn't timeout.  After 8.5 hours it had downloaded around 55 dives which covered all the missing dives and then some more.  I cancelled the download and was given a list of all 55 dives it had downloaded.  I selected all the dives that were missing from my dive log and pressed OK.  All but three of them were loaded.

For some reason, the dives from 22, 26 and 29 December were not loaded into my dive log.  Why would that be?  I have the subsurface log and the libdivecomputer log  Where should I send them?

[Jef Driesen]

On 27/05/2020 23:41, Andrew Trevor-Jones wrote:
> I reran the download attempt with Subsurface creating a log file as well.
>
> This time it didn't timeout.  After 8.5 hours it had downloaded around 55 dives
> which covered all the missing dives and then some more.  I cancelled the
> download and was given a list of all 55 dives it had downloaded.  I selected all
> the dives that were missing from my dive log and pressed OK.  All but three of
> them were loaded.

Hmm, that's really slow. BLE is quite slow, but I wouldn't expect that slow.

> For some reason, the dives from 22, 26 and 29 December were not loaded into my
> dive log.  Why would that be?  I have the subsurface log and the libdivecomputer
> log  Where should I send them?

You can email them directly to me.

Jef

[Andrew Trevor-Jones]

On Thursday, 28 May 2020 17:33:08 UTC+10, Jef Driesen wrote:

On 27/05/2020 23:41, Andrew Trevor-Jones wrote:
> I reran the download attempt with Subsurface creating a log file as well.
>
> This time it didn't timeout.  After 8.5 hours it had downloaded around 55 dives
> which covered all the missing dives and then some more.  I cancelled the
> download and was given a list of all 55 dives it had downloaded.  I selected all
> the dives that were missing from my dive log and pressed OK.  All but three of
> them were loaded.

Hmm, that's really slow. BLE is quite slow, but I wouldn't expect that slow.

 

Long dives (~2hrs each) with log rate at 5 seconds.

 

> For some reason, the dives from 22, 26 and 29 December were not loaded into my
> dive log.  Why would that be?  I have the subsurface log and the libdivecomputer
> log  Where should I send them?

You can email them directly to me.

Sent. Thanks! 

[Andrew Trevor-Jones]

Is there any update on this?  

I still can't load those three dives.

Is there another way I can load them?

[Jef Driesen]

Did you already try my suggestion to use the latest subsurface version? I could
"download" your data without any problems.

Jef