Comment 435 for bug 263435

(In reply to comment #420)
> (In reply to comment #415)
> > - Web applications should have the ability to receive any key, including
> > modified and even function keys, since only then the long-term dream of web
> > apps replacing desktop ones can be achieved;
>
> Although you amend this a little further on I just want to point out that this
> is exactly opposite to the experience gained from multitasking operating
> systems in the last decades. If the plugin/web-app/etc is to behave w.r.t. the
> browser like an application does w.r.t. the operating system, events and
> resources should trickle or be requested from the browser so that they do not
> interfere with the resources of other apps or the events that are expected to
> be handled by an encompassing level (such as kernel or browser). A keystroke is
> such an event. Just like you expect your OS to act in a specific way to e.g.
> ctrl-alt-del or alt-tab there are also browser bindings of this kind,
> especially when it comes to switching focus or closing a tab or window. Right
> now there is simply no way to close the tab or switch to another one with the
> keyboard if a plugin has focus. That is definitely not the correct thing to do,
> whatever philosophy, architecture or dream we intend to follow.

Ok I was a little unclear but I wrote about this later in that comment #415: There's no objection on that shortcuts should be divided between browser and web apps. It would be the best solution. I just thought that the approach in that specific patch is unfair.

I did not intend for web apps to receive literally ALL keystrokes, but to be able to receive all TYPES of keystrokes, including modified (if not already used by browser).

The problem here is that the keys set used by browser might differ per browser, and especially with Vimperator-like addons, so web apps' keys might be blocked, and so it's better to provide means to rebind them to available ones.