As Marius Greuel wrote: > IMHO, we should drop support for parallel ports. I just cannot imagine > anyone still using parallel ports for AVR programming.
As long as it's working, I wouldn't really want to drop it. We abstracted the "logical" layer from the physical one long time ago, and thus the logical layer is being re-used above other physical layers as well. What appears to be reasonable though is to turn the --enable-parport configure default from "yes" to "no", so only those who really want to use the feature get it (similar to linuxgpio and linuxspi). > I wouldn't object so much against using parallel ports if it would not > require a custom driver on Windows. Blame Microsoft for it. All other operating systems (Linux, *BSD, Solaris) offer(ed) a generic parport access driver where the user didn't have to hack the OS for (they only needed to take care about access permissions). Same with USB: all the non-Windows systems offer access to a lowlevel driver layer that can be used without any further configuration (again, except for access permissions) even if there is no other device-specific driver installed. If there are device-specific drivers (e.g. for FTDI), an application can either access that one or the generic one below, without any admin interaction, additional driver selection and so on. Only on Windows, you need hacks like Zadig in order to assign a device explicitly to libusb or other drivers. -- cheers, Joerg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ Never trust an operating system you don't have sources for. ;-)
