git bisect vs libdivecomputer interface

[Anton Lundin]

On 10 January, 2017 - Robert Helling wrote:

> Hi,
> 
> I am trying to bisect the SAC calculation problem and running into problems as the builds of old version fail due to changes in the libdivecomputer API. I need to go to the libdivecomputer source and manually check out old versions there (and due to lack of time I don’t even manage to do that because I get configure errors what I don’t understand when building older versions).
> 
> Is there a way to automate this?
> 
> Do we at least have a list of commits (or tags or whatever) of both libdivecomputer and subsurface to find the points where the API changes (so I can first try those so that git bisect subsurface does not jump between libdivecomputer releases)?
> 
> Any ideas appreciated. I am running out of time for the moment.
> 

I don't have a solution to your libdivecomputer issues, but other
goodies.

I saw the issue you created and got a idea, and found a probable
explanation for the issue.

It looks to be the pressure / volume calculations work done by Linus who
"broke" the sac-tests. The sac-tests are planned for a ideal gas, but
now when we do better sac calculations, we're off by a bunch.

The proper fix would probably to fix pressure-track-code to use real
volumes and not ideal gases, and also fix the sac-test.xml pressures to
match with reality.


//Anton


-- 
Anton Lundin	+46702-161604