[Spice-devel] [spice-gtk v2] Spice-widget: Use new gdk_event_get_scancode if available

[Frediano Ziglio]

> 
> On Mon, Apr 25, 2016 at 04:03:47PM +0100, Frediano Ziglio wrote:
> > This new Gdk API allows in Windows to retrieve the raw scancodes sent
> > by Windows.
> > This allows us to do some translations getting the right value without
> > many hacks and supporting all possible layouts.
> 
> For future reference, can you add some examples of situations where we
> had issues with MapVirtualKey, and where having the raw scancode helps?
> 
> > 
> > Signed-off-by: Frediano Ziglio <fziglio at redhat.com>
> > ---
> >  configure.ac       |  8 ++++++++
> >  src/spice-widget.c | 22 +++++++++++++++++++---
> >  2 files changed, 27 insertions(+), 3 deletions(-)
> > 
> > Changes from v1:
> > - added configure.ac code so use using Gdk from
> >   master can use it.
> > 
> > diff --git a/configure.ac b/configure.ac
> > index ce80d88..ab7385f 100644
> > --- a/configure.ac
> > +++ b/configure.ac
> > @@ -142,6 +142,14 @@ AC_SUBST(GTK_CFLAGS)
> >  AC_SUBST(GTK_LIBS)
> >  SPICE_GTK_REQUIRES="${SPICE_GTK_REQUIRES} gtk+-3.0 >= $GTK_REQUIRED"
> >  
> > +old_LIBS="$LIBS"
> > +old_CFLAGS="$CFLAGS"
> > +CFLAGS="$CFLAGS $GTK_CFLAGS"
> > +LIBS="$LIBS $GTK_LIBS"
> > +AC_CHECK_FUNCS(gdk_event_get_scancode)
> > +LIBS="$old_LIBS"
> > +CFLAGS="$old_CFLAGS"
> > +
> >  PKG_CHECK_EXISTS([gtk+-quartz-$with_gtk], [have_quartz=yes],
> >  [have_quartz=no])
> >  AS_IF([test "x$have_quartz" = "xyes"], [AC_DEFINE([HAVE_QUARTZ], 1, [Have
> >  Quartz?])])
> >  AM_CONDITIONAL([WITH_DISPLAY_QUARTZ], [test "x$have_quartz" = "xyes"])
> > diff --git a/src/spice-widget.c b/src/spice-widget.c
> > index b458909..d666642 100644
> > --- a/src/spice-widget.c
> > +++ b/src/spice-widget.c
> > @@ -1455,9 +1455,25 @@ static gboolean key_event(GtkWidget *widget,
> > GdkEventKey *key)
> >      scancode = vnc_display_keymap_gdk2xtkbd(d->keycode_map,
> >      d->keycode_maplen,
> >                                              key->hardware_keycode);
> >  #ifdef G_OS_WIN32
> > -    native_scancode = MapVirtualKey(key->hardware_keycode,
> > MAPVK_VK_TO_VSC);
> > -    /* MapVirtualKey doesn't return scancode with needed higher byte */
> > -    scancode = native_scancode | (scancode & 0xff00);
> > +    /* Try to get scancode with gdk_event_get_scancode.
> > +     * This API is available from 3.22 or if backported.
> > +     */
> > +#if HAVE_GDK_EVENT_GET_SCANCODE || GTK_CHECK_VERSION (3, 22, 0)
> 
> You don't need the version check with HAVE_GDK_EVENT_GET_SCANCODE (but
> you could add a comment saying this was added in 3.22 to the
> configure.ac check).
> 
> > +    native_scancode = gdk_event_get_scancode((GdkEvent *) key);
> > +    if (native_scancode) {
> > +        scancode = native_scancode & 0x1ff;
> > +        /* Windows always set extended attribute for these keys */
> > +        if (scancode == (0x100|DIK_NUMLOCK) || scancode ==
> > (0x100|DIK_RSHIFT))
> > +            scancode &= 0xff;
> > +    }
> 
> Should we have a log when native_scancode is NULL when we use the
> gdk_event_get_scancode codepath? A bit sad that even with get_scancode
> we end up needing to special case some keys :(
> 
> Christophe
> 

This would only happen if the event is generated and the information
is not filled in.
I'll post a new version.

Frediano