Vasily Khoruzhick wrote:
> Yep, I really don't know what to do with devices that don't support 
> identification :( So now pam_fprint creates full gallery, and if device 
> doesn't support identification, uses only first fiingerprint for verify.

OK, maybe that's the way to go. If you're happy that you can't do more 
with the patch I'll look over it in more detail.

> I've heard that you're writing authentification daemon, or something like 
> that, if it's so, when it will be available for public? I'm asking because 
> now I'm encountering problems with blocking nature of fp_identify and 
> fp_verify functions (I can detail it if you want ;)). Maybe you've seen a 
> part of discussion with Marcin about that on maillist...

I know exactly the kinds of problems you are facing, and have spent 
months working on fixing them :)

Today I reached a milestone - libfprint is now fully asynchronous (well, 
I temporarily disabled drivers for devices I don't own, but have ported 
the others) and provides an asynchronous API to library users. 
fprint_demo is also ported to the asynchronous API.

If you want to take a look you need latest libusb git (master), latest 
libfprint git (libusb-1.0 branch), and latest fprint_demo git (async 
branch).

The prerequisites for starting the authentication daemon were all the 
above tasks, which are now done as of today. So I started working on the 
daemon a few hours ago. I will publish it as soon as there is something 
worth publishing.

This async work is all very volatile. For example the libusb-1.0 API is 
not stable and has work to be done. I have new ideas how I want the next 
incarnation of the libfprint async API to look too (current one will be 
modified without warning). Then the dbus API will hit the same 
instability for a while. I'm in a bit of a hurry preparing for my 
assessed demonstration, but once that is out of the way I can focus 
properly on stabilising things one at a time, starting with libusb.

Daniel
_______________________________________________
fprint mailing list
[email protected]
http://lists.reactivated.net/mailman/listinfo/fprint

Reply via email to