<div class="gmail_quote">On Thu, Feb 23, 2012 at 3:54 PM, Kristian Hoegsberg <<a href="mailto:hoegsberg@gmail.com">hoegsberg@gmail.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> <div class="im">On Wed, Feb 22, 2012 at 02:21:41PM -0700, Scott Moreau wrote: > Ideally, we would want to use <modifier>+Scroll binding but that will have > to wait for axis events. For now we just use keybindings. Zoom in/out with > Super+Up/Down. </div>Applied with a minor edit to really fix the code-before-declaration thing. With your change we still end up with weston_output_update_zoom(struct weston_output *output, int x, int y) { <div class="im"> if (output->zoom.level <= 0) </div> return; float ratio; ... That is, the ratio variable is declared after the if statement. That's what we don't want. </blockquote><div> Ah ok. So all declarations should come before code. Clearly I didn't pay attention to the meaning of code-before-declaration. </div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> Anyway, enough nit-picking, it works well where, but we need to fix the binding issue. We need to swallow the key press that triggers the binding and we also need to swallow the corresponding key up. What I've been thinking is that we could let the binding handler return a value to indicate whether or not to swallow the event. The tricky part is going to be swallowing the release event, since that may come many key events later. We could just maintain a per-device list of "keys that are down, and we need to swallow the release event", but that's kinda gross. I don't really think there's a nicer way though. And all this applies to pointer buttons as well. Kristian </blockquote></div> Yes, this is a problem with all keybindings currently. I was thinking the same about eating the actual keybinding when the pressed/clicked but I hadn't given much thought about the corresponding release event(s). It does seem like we'd need some kind of state tracking for this. I haven't looked into it yet but maybe when a binding is detected, we could set something (possibly on the binding list or similar?) to listen for release of the events involved. I haven't given much thought to this nor have I looked into it yet but will do so. Also, is this a blocker? I don't see the patch in weston yet. Thanks, Scott