GSoC Project: Native Bluetooth support for Linux and/or Windows

[Dirk Hohndel]

> On Mar 9, 2015, at 7:10 AM, Lubomir I. Ivanov <neolit123 at gmail.com> wrote:
> 
> On 9 March 2015 at 15:36, Jef Driesen <jef at libdivecomputer.org> wrote:
>> On 2015-03-09 14:05, Tommi Saviranta wrote:
>>> 
>>> On Mon, Mar 09, 2015 at 15:33:24 +0300, Purity Musyoki wrote:
>>>> 
>>>> In the ideas page, it is noted that there is some sample code of what
>>>> has been done so far regarding the [native] bluetooth idea. I have
>>>> gone through the source code for the Subsurface project, but I am
>>>> unable to find the bluetooth downloader code.
>>> 
>>> 
>>> Perhaps you're looking for the following email and attached patches:
>>> 
>>> http://libdivecomputer.org/pipermail/devel/2014-May/000285.html
>>> 
>>> With those patches libdivecomputer works nicely with my Petrel.
>>> Subsurface GUI also works - even if it requires some persuation.
>> 
>> 
>> I created a libdivecomputer ticket and attached the latest patches there.
>> The last patch in the email (#0006) is completely outdated.
>> 
>> http://trac.libdivecomputer.org/ticket/14
>> 
>> Basically I have a working proof of concept (including the device
>> discovery), for Linux and Windows. What's still missing is the Mac OS X
>> support, and the proper integration in libdivecomputer. The latter is
>> blocked on the libdivecomputer api redesign.
>> 
> 
> so this is an interesting situation,
> 
> if Jeff has a working POC in libdc for Linux and Windows, do we really
> need to reinvent the wheel in Subsurface or should Subsurface use
> libdc for the device discovery and should students contribute in both
> libdc (backend) and Subsurface (frontend if needed) in that way?
> Dirk, how should that be approached GSoC wise?

Definitely a combination of libdivecomputer and Subsurface work.
But it’s more complicated than just using Jeff’s code. It needs a different
user interaction on the Subsurface side - discovery and all that. And when
I last looked at Jef’s code it didn’t work for one of my use cases (but that
may no longer be true).

> also, seems that we need someone with OSX knowledge on the subject.

I am perfectly happy to have a Linux and Windows implementation the
first time around (that’s 80+% of our users). I know that Nick has Mac
Bluetooth code and he mentioned that he could be talked into donating
that to libdivecomputer, so let’s focus on the rest of what needs to be
done…

/D