Jan Kiszka <jan.kis...@web.de> writes: > On 2013-02-25 16:44, Jan Kiszka wrote: >> On 2013-02-25 16:39, Anthony Liguori wrote: >>> Jan Kiszka <jan.kis...@siemens.com> writes: >>> >>>> This is in fact very simply: When the input in grabbed, everything >>>> should be exclusively passed to the guest - except it has our magic >>>> CTRL-ALT modifier set. Then let GTK filter out those accels that are in >>>> use. When checking the modifier state, we just need to filter out NUM >>>> and CAPS lock. >>> >>> Can you explain what you're fixing? >> >> That it's not filtering what it is supposed to. >> >>> >>> We shouldn't hard code modifiers like this. The reason you give >>> accelerators paths like this is so that they can be overridden by a >>> user. >>> >>> That's why I filtered by path. Once we're running, we shouldn't assume >>> that accelerators use the modifiers we started with. >> >> Your path-based filtering does not work as it uses an unsupported >> internal function (see my other mail). >> >> We can make the modifier configurable via QEMU means (command line >> parameter, gconfig, whatever). But let's get the basics working first. > > The bug still exists, my patch still applies. Unless you have some idea > for a better solution, please apply this for now so that CTRL-q inside a > guest doesn't kill more kittens.
Hi Jan, Your patch breaks overriding accelerators which as I said before is a critical accessibility feature. The current code works for me but I realize it's using an unsupported interface. I'll spend some time today trying to find a work around. But we definitely cannot assume that the accelerators are using any specific modifiers. Regards, Anthony Liguori > > Jan