[maemo-developers] [maemo-developers] VirtualKeyboard service/API

[From: Antonio Gomes]

Some keys are not recognized, though. Actually, mainly those related to
deleting actions (like delete, backspace virtual keyboard keys and so on).
As far as I can see/understand, xterm also had to handle them in a proper
way, using "vte" [1] for that. Is it ? Is there other GtkIMContext singal to
handle actions like these ?

[1]
http://developer.gnome.org/doc/API/2.0/vte/vteterminal.html#VTETERMINALERASEBINDING

regards

On 1/28/06, Antonio Gomes <tonikitoo at gmail.com> wrote:
>
> hey Aaron ..
>
> just to thank you for your input regarding the Vk stuff ;)
>
> is works ! (just fixing bugies now)
>
> cheers
>
> On 1/24/06, Aaron Levinson <alevinsn at aracnet.com> wrote:
> >
> > I should note that the "commit" signal, along with the "preedit-changed"
> >
> > signal, are the only ways that you can respond to input from an input
> > method.  That's likely your problem--you weren't doing anything with the
> > string in the "commit" callback.  Writing a proper "commit" callback is
> > a
> > requirement for using a GTK input method.  I wouldn't bother with using
> > gdk_window_foreign_new() if this works, and it should work.
> >
> > Aaron
> >
> > On Tue, 24 Jan 2006, Antonio Gomes wrote:
> >
> > > On 1/24/06, Aaron Levinson < alevinsn at aracnet.com> wrote:
> > > >
> > > > It's unclear to me why you wouldn't be able to "type" anymore in
> > HTML
> > > > forms.  Do you mean with a regular keyboard or with the virtual
> > keyboard?
> > >
> > >
> > >
> > > it stop working only for the virtual keyboard. Regular kb is ok !
> > >
> > > As you indicated in a previous e-mail, you are getting to the "commit"
> > > > signal--why don't you just push the string passed to the "commit"
> > callback
> > > > directly into the form?
> > >
> > >
> > > Hummm ... I got your point here. Good enough to give it a try ;)
> > >
> > > Look at the source code for GtkSocket or maybe GtkPlug for examples of
> > > > using gdk_window_foreign_new().  GtkMozEmbed may use it as well.
> > >
> > >
> > > Ok ... I'll do that now .
> > >
> > > many thanks again
> > >
> > > Aaron
> > > >
> > > > On Tue, 24 Jan 2006, Antonio Gomes wrote:
> > > >
> > > > > > I still think working with the GtkMozEmbed is the way to go--I
> > don't
> > > > > > exactly understand your description of the problem that occurred
> > when
> > > > you
> > > > > > used a GtkIMContext with your GtkMozEmbed widget.  A more
> > detailed
> > > > > > description would help.
> > > > >
> > > > >
> > > > > Ok, so let me try: GtkMozEmbed widget wraps some gtk stuff, making
> > > > easier
> > > > > "embedding" mozilla in a gtk app. It also wraps some mozilla guts
> > (like
> > > > the
> > > > > EmbedPrivate, and listeners - you don't care, probably) and these
> > are
> > > > left
> > > > > alone - I mean, non-gtk widgets. Having said that, the point is:
> > "HTML
> > > > > forms" are not gtk/gdk objects, so I can not connect im_context to
> > them
> > > > > without mapping them to gtk object (I'm investigating ways to make
> > this
> > > > > mapping possible, but I am not sure if it is supported).
> > > > >
> > > > > In other words, the GtkMozEmbed widget itself seems to be too
> > external
> > > > for
> > > > > my needs, I guess. As soon as set
> > > > >
> > > > > *gtk_im_context_set_client_window ((GtkIMContext *)
> > embed->im_context,
> > > > > GDK_WINDOW (widget->window));
> > > > >
> > > > > *
> > > > >
> > > > > I can not longer type into HTML forms ... Now, does it make sense
> > to you
> > > > ?
> > > > >
> > > > >
> > > > > Why don't you send me the relevant source code
> > > > > > that you wrote for GtkMozEmbed?  I'll have a chance to look at
> > it over
> > > > the
> > > > > > weekend or possibly on Friday.
> > > > >
> > > > >
> > > > > I will ... thanks for that ;)
> > > > >
> > > > > Regarding gdk_window_foreign_new, the GdkNativeWindow parameter
> > (not
> > > > > > GtkNativeWindow as you stated) represents a native window.  As
> > stated
> > > > in
> > > > > > the GDK API documentation, this is a Window for X and an HWND
> > for
> > > > Win32.
> > > > > > See
> > > > > >
> > > > > >
> > > >
> > http://developer.gnome.org/doc/API/2.0/gdk/gdk-X-Window-System-Interaction.html#gdk-window-foreign-new
> > > > > > for more information.  I've never used this function and I
> > haven't
> > > > > > personally tried to use the method that I suggested in my
> > original
> > > > e-mail.
> > > > >
> > > > >
> > > > > it seems nobody knows much about it :S
> > > > >
> > > > > regards
> > > > >
> > > > >
> > > > > > > Aaron,
> > > > > > >
> > > > > > > After some troubles, I'm now reconsidering the idea of
> > creating
> > > > > > > GdkNativeWindow from the HTML form, and setting the im_context
> > to it
> > > > > > > (gtk_im_context_set_client_window) - as your proposed first -
> > > > instead of
> > > > > > > setting it to the GtkMozEmbed widget !
> > > > > > >
> > > > > > > But I'm stuck on the gdk_window_foreign_new function. How do I
> > use
> > > > that
> > > > > > ? I
> > > > > > > mean, I don't get the GtkNativeWindow parameter required by
> > this
> > > > > > function
> > > > > > > ... Could you help me out here ?
> > > > > > >
> > > > > > > thanks again ;)
> > > > > > >
> > > > > > >
> > > > > > > On 1/23/06, Antonio Gomes < tonikitoo at gmail.com> wrote:
> > > > > > > >
> > > > > > > > hi Aaron,
> > > > > > > >
> > > > > > > >
> > > > > > > > > I didn't mention this in my last e-mail, but you'll also
> > need to
> > > > > > setup a
> > > > > > > > >
> > > > > > > > > signal handler for the GtkIMContext object's "commit"
> > signal.
> > > > > > > >
> > > > > > > >
> > > > > > > > I have it (the 'commit' signal) already connected and also
> > its
> > > > > > callback
> > > > > > > > defined as it is in your post at
> > > > > > > >
> > > > http://maemo.org/pipermail/maemo-developers/2006-January/002337.html-
> > > > > > > > cool !!! ...  But problem begins after creating the
> > IM_Context
> > > > object
> > > > > > and
> > > > > > > > associating it with the GdkWindow of my GtkMozEmbed object :
> > the
> > > > > > > > mozilla_engine stop recognizing any input from the vkb,
> > except if
> > > > I
> > > > > > press
> > > > > > > > 'backspace' ...
> > > > > > > >
> > > > > > > > bwt, in 'commit_cb' method, it prints what I type at the
> > console
> > > > but
> > > > > > > > doesn't add nothing at the HTML form (detail: it does with
> > no
> > > > context
> > > > > > > > associated or if I type from the physical keyboard) ...
> > > > > > > >
> > > > > > > > static  gboolean
> > > > > > > > gtk_moz_embed_im_context_commit_cb (GtkIMContext *ctx,
> > > > const  gchar
> > > > > > *str, GtkMozEmbed *aEmbed) {
> > > > > > > >   gunichar uChar = 0;
> > > > > > > >   guint keyval = 0;
> > > > > > > >   gchar *text2 = (gchar *) str;
> > > > > > > >
> > > > > > > >   printf ("\n [GTK_MOZ_EMBED] it gets here '%s'
> > > > > > > > \n", str);
> > > > > > > >
> > > > > > > >   *// the string is in UTF-8 format--iterate through the
> > > > characters*
> > > > > > > >   *while*  (*text2) {
> > > > > > > >     uChar = g_utf8_get_char(text2);
> > > > > > > >     keyval = gdk_unicode_to_keyval(uChar);
> > > > > > > >
> > > > > > > >     *//*
> > > > > > > >     *// do something with the character here*
> > > > > > > >     *//*
> > > > > > > >     text2 = g_utf8_next_char(text2);
> > > > > > > >   }
> > > > > > > > }
> >
>
--
--Antonio Gomes
http:// tonikitoo (dot) blogspot (dot) com (slash)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.maemo.org/pipermail/maemo-developers/attachments/20060130/bba360b7/attachment.htm