possible O2 PP display bug

[Linus Torvalds]

On Sat, Mar 2, 2013 at 1:18 PM, Jan Schubert <Jan.Schubert at gmx.li> wrote:
>
> It's 1,04, see the same dive freshly downloaded using current git attached.
> Dirk, this also proves 3.0.2 is doing well, checked all the dives
> available on my Shearwater.

Ok, good. This also works as a good example of how trying to fix
things up isn't all that obvious. For this particular value, 1.04 was
the closer value in this case even if it was border-line, but for
things like 1.01-vs-1.10 or 1.02-vs-1.20 it would have been basically
impossible to guess.  So the "we can try to fix things up" would at
most have removed the obvious errors, not all errors.

The "we can guarantee it was saved only once" case could have solved
it, but it looks to me like your dive computer fundamentally returns
data with "centibar" accuracy, so you'd actually never have the
particular values (like 1.008 which would be saved as "1. 8" with a
space) that would have been markers of "this has been saved only
once". So we'd never have a case of really *knowing* that the data was
saved only once, and guessing would be our only choice.

So re-downloading is, I think, undeniably the right thing to do.

             Linus