On Tue, Mar 18, 2008 at 2:07 PM, ?tienne Bersac <bersace03 at gmail.com> wrote: > Hi, > > > > extending sane_open just a bit to always > > support 'libusb:xxx:yyy' or '/dev/sgX' is not an API change, so could > > be implemented much sooner, and really only requires changes to the > > backends that dont already use those names, which is few. > > Right. Considering that HAL currently support usb, scsi and ieee1394, it > makes sens to have a common scanner naming policy : > > * USB = "<backend>:libusb:<busnum>:<devnum>" > * SCSI = "<backend>:<devfile>" > * IEEE1394 = "<backend>:?" > > I have no idea for IEEE1394, but it would be nice to provide also a well > known device name for IEEE1394. > > With a well known naming policy, a HAL callout could just compute the > sane device name on plug and publish it as HAL device property. > > Once we have a resonnable policy for well known device name, we could > list backend that needs to comply. I didn't find, but i'm not very > skilled at search in all sane backends. > > Regards, > ?tienne.
this requirement for naming will never fly. i have a patch for the fujitsu backend right now that will use 'fujitsu:model:serial' for those machines that can give me their serial. this is needed for users who have two scanners connected to same machine and want to uniquely identify them from command line. and no- hal cannot get the model or serial without asking the backend, they are not in the usb descriptors. i think the best you can do is to require that the backend accept alternate names for the scanner if it is going to use something other than the list you gave above. oh- and firewire scanners show up on linux as /dev/sg0 when they use the sbp2 (scsi over firewire) module. I dont know about others OSes allan -- "The truth is an offense, but not a sin"
