[Subsurface-divelog/subsurface] Merged dives lose gas changes (#364)

[Linus Torvalds]

On Sat, Apr 29, 2017 at 3:52 PM, Linus Torvalds
<torvalds at linux-foundation.org> wrote:
>
> I suspect that this kind of merging (that *concatenates* two different
> dives into one, rather than merging two different dive computers' same
> dive into one dive with two computers) probably never merged events at
> all in the first place.
>
> Because it's not just gas changes that are missing, all the other
> events are gone too.

No, we have the whole "merge_events()" code that does this.

But commit b5de08b7 ("No gas change event on merging dives with same
gas") is completely bogus. Changing the 'p' pointer (which is the
"this is where we'll insert the next event we pick") to point to one
of the source list entryes is just broken.

I think that needs to be reverted, but I'll look at this a bit more.
Maybe I can just fix it up.

                          Linus