Incorrect tank info.

[Linus Torvalds]

On Tue, Aug 21, 2012 at 1:28 PM, Ďoďo <dodo.sk at gmail.com> wrote:
>
>  1. I opened XML I created some months ago.
>  2. Then I imported new data from D6.
>  3. And then I noticed that all my dives I dived with air had O2 31%.
>  The dive I really had Nitrox 33 had correct 33% and later dive with
>  Air got already Air.

Can you add some debugging to libdivecomputer.c:parse_gasmixes().

Just a simple

    fprintf(stderr, "o2=%d he=%d\n", o2, he);

after the two lines that gets it from the libdivecomputer gasmix field.

IOW, the code should look something like

                o2 = gasmix.oxygen * 1000 + 0.5;
                he = gasmix.helium * 1000 + 0.5;
                fprintf(stderr, "o2=%d he=%d\n", o2, he);

>  6. It looks like it was taken from older dives.

I really don't think we should be leaking data from older dives when
importing, because we create the new dives with "alloc_dive()", which
does a memset() to clear out all fields.

But I'll see if I can reproduce odd gasmixes with my setup.

>  8. Then I imported all to data without reading older XML and the same
>  problem. My last trip took the cylinder content from the trip before.
>  Trip before took from 2 trips before. Till the air was not found.
>  9. Interesting is, that I did not observe it always. In the past there
>  are dives when trip finished with Nitrox 36 but next dive is correctly
>  with air.
>
>  Any idea what can be the source of the problem?

My guess would be libdivecomputer changes. Maybe a D6-specific issue?

Especially if it happens without even reading the older XML file.

              Linus