regressions because of the new tank handling

[Linus Torvalds]

On Wed, Sep 13, 2017 at 11:20 AM, Dirk Hohndel <dirk at hohndel.org> wrote:
> Given that I have been rather distracted for the last few weeks, I haven't
> followed up on making sure that the changes to the tank handling that
> Linus made a while ago didn't have any other unintended side effects.

The tankbar issue was not an "unintended side effect". I very much
talked about it, and you said you'd fix it up because it was your
code.

There's a helper "get_gasmix()" now, with a few example uses, but it
basically boils down to

        struct gasmix *gasmix = NULL;
        struct event *ev = NULL;

        iterate forward over time {

                gasmix = get_gasmix(dive, dc, entry->sec, &ev, gasmix);

        }

because you can not look up the current gasmix by just looking at the
plot-info "sensor" any more.

> It appears that Stefan found a couple which he reported here:

Hmm. The infobox pressure thing I will look into.

The spikes in the data I would need a test file for to see what's up.

              Linus