Daniel Drake wrote:
>> Yes, this is for that reason. I think I'll move it to the loop sooner
>> or later. For now, I leave it here as I'm not sure about the number
>> of strips we should capture between every gain adjustement. I'll fix
>> it with gain stuff. I think.
>
> Looking at the code closer, I can see why you've done it this way: you
> do 2 initial captures before sending strip_scan_reqs. Maybe these
> initial 2 scans are used by the windows driver for gain calibration
> and should be discarded from the resultant image...
I've always considered these first captures as normal. I didn't thought
about some sort of calibration strips... I'll look into it.
>>>> + // img->height must be >= 12 (if not, we get seg fault)
>>>> + // img->height = 8 when we don't move the finger on the sensor
>>>> + // We set it to 12 to avoid a crash.
>>>> + if (img->height < 12)
>>>> + img->height = 12;
>>>> +
>>>
>>> Why 12? Where does the segfault happen?
>> 12 because this is the minimum to avoid the segfault. This is a dirty
>> fix! The segfault happens on line 132 of libfprint/nbis/mindtct/maps.c.
>
> OK. We should fix this properly, so I left this part out of the patch.
> I wonder if NBIS simply fails on tiny images. If you think that sounds
> like a possible explanation, could you please file a bug for this at
> http://www.reactivated.net/fprint/bugs/
Done. I've made a mistake on the line where the segfault happens. The
correct one is in the bug report.
> The rest of the patch is committed locally.
>
> Thanks!
> Daniel
>
_______________________________________________
fprint mailing list
[email protected]
http://lists.reactivated.net/mailman/listinfo/fprint
