On Jan 13, 2009, at 5:48 PM, Peter Stuge wrote: > Charles Lepple wrote: >> If you send me some pointers to the "raw hid" API, I will take a >> look at it. I am in the process of rewriting the libhid backend to >> not be tied to just raw libusb calls. > > /usr/src/linux/Documentation/usb/hiddev.txt > > I would like to request that the backend becomes a separate library > (almost) as simple as the Windows HID API. libhidio may be a good > name? I don't know how much sense it really makes, but I think it > would be nice to have.
Peter, I'm not sure I understand your suggestion here. What are you proposing to separate? When I say "backend", I mean the glue code to a platform's native HID API (or to libusb, for "legacy" libhid operation). The "frontend", in my mind, is the generic API that hides platform-specific details. I know that the Windows HID API automates finding the first input and output report, and finding which endpoint to use, but to me that sounds like a maintenance nightmare, with the new #1 FAQ becoming "why doesn't my Windows code work on Linux?". To me, that also sounds like a convenience function on top of the frontend described in the previous paragraph (for platforms other than Windows), since it builds on the more generic libhid API. Do you have a sketch of your proposed architecture? _______________________________________________ libhid-discuss mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/libhid-discuss http://libhid.alioth.debian.org/
