ABC <abc at telekom.ru> wrote: (resending to list)
Hi, >> I've just tested xerox_mfp on the 1815dn, and it works very well, >> although I've had two segfaults with XSane. It could be related to >> SANE 1.1, though. A color scan from the ADF will segfault midway; >> although the backtrace seems to indicate it's an XSane issue, it also >> looks like the stack is broken. > > You may send valgrind log file to me for analysis. I'm not sure the backend is at fault. scanimage doesn't segfault like XSane does, which would indicate a problem in XSane. But I'll look into it. >> Other than that, the backend is conflicting with usblp to claim the >> USB device; do you have a way to fix that? Would going over the >> emulated parallel port be possible? > > Strangely, as I'm using standard SANE interface to communicate with usb > device. About parallel port - my device is only using usb, so I really > don't know much to help about parallel ports. The Samsung driver goes over the parallel port, so if the machines are the same, the xerox should support that mode too, I guess? >> I don't really care as I'm printing over the network anyway and use >> USB only for scanning, so I can always blacklist usblp, but not >> everybody can do that :/ > > Any suggestions from your experience how to resolve that? libusb provides a function to unbind the kernel driver from the device, but then you lose printing if you're printing over USB. Unbinding and rebinding afterwards could be a workaround if rebinding is actually possible, but you'd have to unbind/discover/rebind in sane_get_devices and unbind in sane_open then rebind in sane_close. This will also "break" copy scripts and XSane's copy mode as the printer won't be available while the scanner is open (as in sane_open). This all look pretty suboptimal :/ MFP have always been an issue :( JB. -- Julien BLACHE <http://www.jblache.org> <jb at jblache.org> GPG KeyID 0xF5D65169