OK, I understand. Thanks!

On Sun, 10 May 2020 at 00:05, Mike Jumper <[email protected]> wrote:

> On Sat, May 9, 2020, 12:36 gabriel sztejnworcel <[email protected]>
> wrote:
>
> > Thanks for you answer Mike!
> >
> > What if the server has 2 layouts, for example French (azerty) and Hebrew,
> > and the user wants to switch between them during the session? Won't that
> be
> > a problem?
> >
>
> Yes. The user should switch layouts on the client side, not within the
> server.
>
> If the client can send the key scancodes, won't that solve the problem in
> > this scenario?
>
>
> It would shift the location of layout switching to the server. Doing so
> would make things behave the way you specifically want, but may be a step
> backwards from the goal of making things as seamless as possible.
>
> Is it technically possible to implement this?
>
>
> With recent changes to JavaScript key events, possibly, but not across the
> board. Not all input devices have scancodes.
>
> The biggest concerns here would be:
>
> 1) Loss of seamlessness. Keyboard behavior *should* be dictated by the
> behavior of the client's keyboard.
>
> 2) Loss of compatibility. Not all platforms will have scancodes, let alone
> ones that correspond to Windows.
>
> 3) Loss of functionality. Not all keys produce key events, and some keys do
> so only unreliably.
>
> It may be reasonable to provide the option at the connection level, to
> allow the administrator to choose whether to favor the client-side layout
> (and translate to a known server layout) or the server-side layout (and
> entirely ignore client-side key identity).
>
> Philosophically, I don't like it. If such a change is implemented, I would
> be concerned that it might be used as a crutch on the admin side (to avoid
> having to think about what layout is selected), to the detriment of users.
>
> It is something that may be necessary in the case that keyboard layout MUST
> be changeable within the remote desktop. Direct connections to VM consoles
> come to mind.
>
> Even if implemented, I don't think that is the solution you should look
> toward. It would be better to allow your users to transparently use their
> keyboard as it is locally configured, without having them worry about
> manually configuring things again within the remote desktop. This is the
> way things are already designed.
>
> - Mike
>

Reply via email to