On 09.01.2018 12:17, Paolo Bonzini wrote: > On 09/01/2018 12:00, Thomas Huth wrote: >> On 09.01.2018 10:58, Paolo Bonzini wrote: >>> On 08/01/2018 08:23, Thomas Huth wrote: >>>>>> But instead of introducing a new "-braille" parameter, maybe we should >>>>>> rather keep some of the convenience "-usbdevice" possibilities around? >>>>>> E.g. keep "-usbdevice braille", "-usbdevice mouse", etc. but remove >>>>>> things like "-usbdevice serial" and "-usbdevice host" where the code is >>>>>> rather ugly and the user do not gain much in comparison to "-device" ? >>>>> That would work too. But I'm not sure why keep all the usbdevice >>>>> infrastructure when we can do the same (at the price of a small cmdline >>>>> incompatibility) with only 50 lines of code. >>>> I'm just afraid that we will end up with more new parameters in the end >>>> than just "-braille" - or do we feel confident enough that "-usbdevice >>>> braille" is the only one that would need a sugared replacement? >>> >>> I agree that we should plan for suboptions, so it could be "-braille >>> [serial|usb]" for now. >> >> I'm rather afraid that someone else will also complain about the removal >> of the other usbdevice options, so that we'll finally end up with new >> -usbhost, -usbbt and -usbyounameit convenience options... in that case >> it might be better to keep "-usbdevice" instead? > > We can and should say no. Sometimes we can also say yes though. :) > > I just think that Braille is worth a special case because a subset of > our user base (blind people) will use it 100% of the time, plus it is > not supported by libvirt and hence virt-manager.
OK, then let's go forward and use your "-braille serial|usb" patch, and nuke the -usbdevice option completely (assuming Gerd agrees, being the USB maintainer). Thomas