Suunto EON Steel breakages..

[Linus Torvalds]

On Mon, Jun 20, 2016 at 1:17 PM, Axel Richter <axel.richter68 at yahoo.com> wrote:
>
> Hi Linus,
>
> I use an EON Steel, too and will update the subsurface sources next week for
> compiling.
> Should I avoid updating libdivecomputer?

So the breakage isn't in the downloaded data itself, it's just that
libdivecomputer now gives the (incomplete and arguably incorrect)
serial numbers in the old broken legacy format, and subsurface as a
rule takes that legacy information over the more modern string-based
one.

So the download itself is fine, but subsurface will think that it's a
copmpletely different Suunto EON Steel, and that you dove with *two*
otherwise identical dive computers. So it will re-download old dives,
and then if you don't uncheck them in the download window before
pressing "Ok", subsurface will now have *two* EON Steel dive computers
for the overlapping dives - with different dive computer ID's (and
different dive ID's).

So the data itself isn't "wrong", it's just that the bogus "devinfo"
data makes subsurface think it's a new dive computer.

So even with a new libdivecomputer library, you can avoid the
confusion by manually checking which dives you already have, and
making sure you only download new dives (by unchecking the boxes for
the dives you already have).

It's mainly an annoyance, and a source of confusion, in other words.
Part of the annoyance is how long the download takes - downloading
every dive takes potentially several minutes (after which you have to
uncheck the old ones), instead of just a few seconds for the last
couple of dives.

I know how to make subsurface ignore them even with the new
libdivecomputer, I've just been out most day. I'm sitting down now to
try to write the right patch.

              Linus