interesting little bug...

[Linus Torvalds]

On Mon, Mar 25, 2013 at 1:33 AM, Robert C. Helling
<robert at euve10195.vserver.de> wrote:
>
> On 25.03.2013, at 04:31, Linus Torvalds <torvalds at linux-foundation.org> wrote:
>
>> Yes. We have absolutely no support for that, and it's nontrivial to do.
>
> Maybe I'm missing something. But don't we assume at some point (e.g. gas change events) that we can refer to cylinders via their content (i.e. O2/He permilles)?

Yes. Except we really would want to support multiple cylinders with
the same gas, although right now the libdivecomputer interfaces don't
really allow for that (because they only give the gasmix for the gas
switch event).

However, the fact that we refer to cylinders by their gasmix does
sadly not mean that the dive computers

 (a) agree on gas mixes (people make mistakes, and some dive computers
are limited in number of gases or mixes supported)

 (b) use the gasmix to refer to the cylinder at all. The gas change
event is the only one that actually uses the gas mix, most dive
computers tend to refer to their gases by "index" in all other
settings (ie when doing cylinder pressure in the samples).

So we have a mixture of "cylinder index" and "gasmix", and both are
deficient (in different ways).

             Linus