Quoting Paul Smedley <p...@smedley.info>: > Hi Gerald, >=20 > On 28/6/2005, "Gerald Murray" <gmur...@cloudnet.com> wrote: >=20 > >Quoting Paul Smedley <p...@smedley.info>: > >> One query I have though is about devname > >>=20 > >> I see in sanei_usb.h that the libusb format for this is > libusb:busno:devno > >>=20 > >> Is this defined for all devices or just ones that sanei_usb_init thi= nks=20 > >> are scanners? > > > >It is defined for each usb device. Having this naming scheme allows t= he=20 > >OS to use the same functionality that applies to filenames for usb dev= ices. > >So the permissions on that filename can be used to indicate the permis= sions > >for its associated usb device.=20 > Well permissions are pretty much non-existent on os/2... >=20 > >> I assume that the format of devname really isn't important - so long= as=20 > >> it's useful to the usb method being used (ie kernel, libusb or usbca= lls=20 > >> for OS/2) in determining which usb device is being attached to? > >>=20 > >> In which case, I could use usbcalls:deviceno or even just deviceno? > > > >For portability, it would be best to use just one system for all usb > devices > >within that operating system. This is a problem for the OS you are us= ing.. > I don't see why it's a problem?
To clarify, I was suggesting that there should be a uniform way of using the devices, and that 'problem' lies more in the OS, and less in the=20 driver author. It eases portability in adapting any driver to another system or in converting programs to any one specific computer. I did not intend to suggest that there is an obstacle for you there. > usbcalls uses a device number for each usb device that's attached to th= e > system. >=20 > Using the format above, I could have for eg. > usbcalls:0 - ie my USB Floppy > usbcalls:1 - ie my USB Printer > usbcalls:2 - ie my USB Scanner >=20 > When usbcalls:2 is passed to sanei_usb_open, I know to open device numb= er > 2 via usbcalls? Or am I missing something? Looks like you are on the right path. regards, Gerald
