[sane-devel] scanner that worked with linux 2.4 does not work with 2.6 : libusb fiasco (maybe firmware upload problem)
Sun, 7 Nov 2004 12:52:13 +0100
> You don't need to, it'll be found automatically.
yes, if I were lucky. But I'm not.
So let's say it "should" be found automatically.
> > sane-find-scanner finds none
> Try running sane-find-scanner without any arguments. If no scanner is
> found it's most likely a permission problem.
This was the first I tried. It finds none.
I have the permissions set up.
Also, just to avoid permission problems,
I'm trying to run as root before I move
to a less powerful (dangerous) user.
Also, running sane-find-scanner -v
would (as far as I understand right) warn immediately
if it was just the permission.
> See "man sane-usb" for possible solutions.
I read it over many times, and tried everything.
> > I played a lot with snapscan.conf, so far best results with:
> > firmware /usr/local/scanner/benq4300u/u176v042.bin
> > usb 0x04a5 0x20b0
> > /proc/bus/usb/001/022
> > libusb:001:022
> You only need two entries, the firmware entry and the "usb 0x04a5
> 0x20b0". You _must not_ use any device entries.
Those I tried first:
usb 0x04a5 0x20b0
(and also in reverse order).
> The problem is that the device file will change dynamically depending
> on when and how often you switch on your scanner. There's no way to
> have the scanner use the same device file every time.
I'm aware of that, I just tried everything since nothing documented worked
with linux kernel 2.6.
I could find the device automatically using a 5-liner script, if that
Let me rephrase. Without the device hint , neither
sane-find-scanner -v (no device argument)
scanimage -L ( just the mentioned 2 lines in snapscan.conf, no device )
/etc/sane.d# scanimage -L
[sanei_debug] Setting debug level of snapscan to 128.
[snapscan] sane_snapscan_init: Snapscan backend version 1.4.17
[snapscan] sane_snapscan_get_devices (0xbffff314, 0)
No scanners were identified. If you were expecting something different,
The desription in the last mail was the best-so-far result, when
the snapscan backend actually tries to talk to the device.
I tried resetting the scanner from time to time (and changed device bus
location to match)
to avoid the scanner being confused of some messages.
I can grant r..t access to the box if it helps
(if someone asks in private mail, mentioning their /src/ host IP if
# dpkg -l 'sane*'
ii sane 1.0.13-1 scanner graphical frontends
ii sane-utils 1.0.14-7 API library for scanners --