dive planner fixes [was: Re: source code may be getting a bit intimidating]

[Linus Torvalds]

Hmm. Playing around with it, I eventually get a lockup. It seems to be
the same tissue-limit thing as before, though, and not really related
to the planning itself. Although re-planning (by editing fields) does
seem to be involved, but it's probably simply that the bug requires
some special deco history or something.

And yeah, "cancel" doesn't work right. I tried just deleting the last
dive, and that caused a SIGSEGV. I think that was because the code
tried to free() the dive computer model (it's set up a constant
string), but I might be completely wrong about that - there might be
something more fundamental going wrong.

              Linus