It's definitely a bug. And you're right, it's still not working in the epson module, even when you change the sscanf from %d to %i, so there's probably another bug somewhere.
However, when I copied the same code into fujitsu.c, changed the sscanf from %d to %i, and on success (sscanf returns 2) I call sanei_usb_attach_matching_devices(...) [instead of sanei_config_attach_matching_devices], I can now configure my fujitsu scanner using "usb 0x04c5 0x1042", and scanimage recognizes the scanner and I am able to scan just fine. My guess is that either sanei_config_attach_matching_devices has a bug, or possibly more likely, it's more appropriate to call sanei_usb_attach_matching_devices, since we know for sure that we're trying to attach a USB device. Till Kamppeter wrote: > Henning Meier-Geinitz wrote: > >> >> I think it's ok. The epson backend checks for "usb vendor product" and >> calls sanei_usb_attach_matching_devices if it find such a line. >> Otherwise it checks for "usb" and calls attach for the device name. If >> it's not "usb", it calls sanei_config_attach_matching_devices(line, >> attach_one); >> >> > > Are you sure that the "epson" backend understands something other than > fixed devices? I have tried > > usb 0x04b8 0x0112 > > for my Epson Perfection 2450 and the "epson" backend did not find it. > The only way to get this scanner working is to use a fixed address as > > usb /dev/usb/scanner1 > > Till > > > _______________________________________________ > Sane-devel mailing list > [email protected] > http://www.mostang.com/mailman/listinfo/sane-devel >
