galileo luna

[Jef Driesen]

On 2013-08-20 15:56, Henrik Brautaset Aronsen wrote:
> On Tue, Aug 20, 2013 at 3:52 PM, William Perry <wmperry at gmail.com> 
> wrote:
> I spent some more time disassembling the JTrak jars that have a full 
> IrDA
> stack in them - need to carve out some more free time to write a
> proof-of-concept to dynamically load those and get a memory dump.
> 
> That would be seriously exiting!  Any findings yet?

I have looked at JTrak too, when I did the IrDA implementation. From 
memory, they are using the open source Java IR library [1] for the IrDA 
part, and some native glue code for the communication with the hardware 
(MosChip MCS7780 and SigmaTel 4210 chipsets).

In theory it's possible to do the same: implement (or re-use in JTrak's 
case) a minimal IrDA stack and interface with the low-level infrared 
hardware directly. But that's a non-trivial task which, as far as I'm 
concerned, is outside the scope of the libdivecomputer project. I simply 
don't have the resources nor the motivation to work on something like 
that.

And to be honest, I consider adding support into libdivecomputer by 
means of loading the java libraries, a really ugly hack. I highly doubt 
I'll ever accept anything like that.

If you're just trying to get a memory dump, then it's probably easier to 
download the data on linux or windows, and then process it on your mac. 
Or, as mentioned in the blog post from William's email [2], use a small 
device like a Raspberry Pi for the downloading part.

[1] http://jir.sourceforge.net/
[2] 
http://patrick.wagstrom.net/weblog/2012/12/04/30-meters-underwater-with-a-dead-physical-layer-protocol/

Jef