On 16 July, 2014 - Dirk Hohndel wrote:

> On Wed, Jul 16, 2014 at 11:22:32PM +0200, Anton Lundin wrote:
> > This started out yesterday with me looking at dives/test20.xml and saw
> > that the gasname text in the info box is wrong there. It uses the
> > gaschange event data via get_gasmix_from_event and then on to gasname.
> > 
> > The problem with using the result from get_gasmix_from_event is that it
> > doesn't care about the difference between SAMPLE_EVENT_GASCHANGE and
> > SAMPLE_EVENT_GASCHANGE2, and thus doesn't map the gaschange event to the
> > cylinder its changing to.
> 
> ??? That sounds odd...
> 
> > I dug around and found the same type of assumption in lots of places
> > allover the code. A while back pre 4.0 somewhere i fixed up
> > get_cylinder_index to do the right thing and switched all the places to
> > using it, but now we have reverted to using event data again.
> 
> I don't think using event data is WRONG per se - as long as it's done
> correctly. After all, that's how we track gas changes.
> 

You can't use the event data straight off right now. Eg.
SAMPLE_EVENT_GASCHANGE only contains the oxygen part of the gas, and in
that case you wouldn't be able to distinguish between air and 21/35.

> > Should we stop supporting SAMPLE_EVENT_GASCHANGE and converting
> > everything into actually relying on having the exact correct gas in the
> > event?
> 
> It should be there. I'm lost why / where you think it isn't.
> I must be missing something.
> 

If your cylinder says 32.1% and your gaschange event says 33%, what
should we say that the diver is breathing?

In the get_cylinder_idx we map that gaschange event to the closest
matching cylinder, aka we say that the diver is breathing 32.1% , if we
use the gaschange event straight off, we would say the diver is
breathing 33%.

The difference is even greater if we look at SAMPLE_EVENT_GASCHANGE
switches. There the event could say 18% where the cylinder mapping says
18/45.

> > The other option is to stop looking at whats in the event and start only
> > looking at what gas is mapped to different times in the dive. We can't
> > do both, and mix and match them.
> 
> Well, post 4.2 I really want to change the way we deal with events, but it
> sounds like this is somewhat independednt from that intended change.
> 
> But I think I need to understand better the actual underlying issue before
> I can make any sane suggestions...
> 

The patch i sent earlier about the tool tip is the only _really_ visible
i found, but there are probably more bugs out there, but they are going
to be a bit more invasive to fix, so the question is, should we go with
mostly/probably right for 4.2 or should we fix them before 4.2?

//Anton

-- 
Anton Lundin    +46702-161604
_______________________________________________
subsurface mailing list
[email protected]
http://lists.hohndel.org/cgi-bin/mailman/listinfo/subsurface

Reply via email to