Gustavo Chain wrote:
> There are minor changes to previously submitted driver.
> 
> Fixes:
> * Setup bz3_threshold to 23 (works much more accurate)
> * Replace USB_ENDPOINT_TYPE_ISOCHRONOUS by USB_RECIP_INTERFACE |
> USB_ENDPOINT_OUT

That's not right constant usage either. USB_RECIP_INTERFACE is a 
constant designed for use in the bmRequestType field of control 
messages, not for endpoint addressing.

Also there is still "constant abuse" in usb_bulk_read() and 
usb_clear_halt() and maybe other places too.

> * Remove awayt_finger_on (not implemented)

Seems to me that this should be on your TODO list, or maybe we should 
provide a generic one in the library which uses a histogram-like 
approach to detecting fingers.

> * .id = 7 (is that correct?)

I'll assign one at merge time, use whatever number you feel like for now.

> TODO:
> * Add -D_GNU_SOURCE to Makefile ( for memmem(3) )

Is there a glib equivalent of memmem()? It seems like requiring 
GNU-specific libc extensions may hurt portability.

You should probably do error checking in bulk_write_safe() callers as 
this may help detect weird bugs early on in future.

Apart from that, it looks great. If you send a new revision with those 
changes I'm happy include it and you can submit the later items as patches.

BTW: in the development branch (fpusb), all drivers have to be converted 
to use async USB I/O which unfortunately adds a little driver complexity 
which I'm working to minimize. I'm happy to do the porting but you will 
need to test the new version at some point :)

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

Reply via email to