* grouped all display init in gdkdisplay-x11.c
* created a per display X ID hashtable
* create per display equivalent of gdk_[window|pixmap|font]_lookup & co.
* added various convienience functions
- revert gdk_x11_*.grab_server changes and introduce gdk_x11_display_*.grab
- add screen property to GtkInvisible
- add g_return_if_fail (GDK_IS_DISPLAY (display)) or equivalent
for all public API taking a GdkDisplay.
- ditto for GdkScreen
- mark gdk_event_send_client_message as not multihead safe and
added gdk_event_send_client_message_for_display
- add gdk_send_xevent_for_display() and make gdk_send_xevent()
as not-multihead-safe.
- move leader_window to GdkDisplay consequently change
gdk_set_sm_client_id_for_screen to for_display
- added GDK_MULTIHEAD_SAFE around all non multihead fcts.
* remove includes of screen/display as Owen had them to gdk.h
* revert gdk_x11_*.grab_server changes and introduce gdk_x11_display_*.grab
* store colormap per screen in gdk_colormap_get_sytem_for_screen
* replace gdk_rgb_ditherable_for_screen() by gdk_rgb_colormap_ditherable()
* add screen property to GtkInvisible
* added idef GDK_MULTIHEAD_SAFE around gdk multihead unsafe fct
* added sanity check in gdkdisplay-x11.c and gdkscreen-x11.c
* gtk/gtkclist.c check if the widget is realized before refreshing it
* gtk/gtkfontsel.c created a realize function to hold all the screen dependent
resources initialization.
now gtk font selector is multihead safe
* fixed multihead support for colorselection dialog
* various cursors are now created on the right display
* popup window are popped up on right display/screen
- revert _gdk_x11_display_impl_get_type to gdk_x11_display_impl_get_type
as its not a local symbol (it's used in gtk and immodules)
- remove gdk_window_set_screen.
- cleaned gdk_window_set/get_mwm_hints from static atoms.
- gdk_window_iconify use proper screen.
- remove assert and change comments in gdk_window_foreign_new_for_display
- gtk_drag_get_ipc_widget uses GdkDisplay now
- reimplemented GdkDrawable get_display() in term of GdkDrawable get_screen()
- rename gdk_display_manager to _gdk_display_manager and remove it from gdkx.h and gdkprivate.h
- removed scr and dpy abbreviations.
strange, selection and dnd seems half broken, e.g. outside processes can't grab pasted
selection. This is a problem introduced in the HEAD not gtk-multihead branch
as it WAS working before merging.
paste works for out process selection to inprocess
still doesn't work for in process to out process (something funny with
gtk_selection_request not invoking the registered selection convertion
handler....).
changed all the gtk multihead design remove screen variable from GtkWidget to GtkWindow
now gtk_widget_get_screen recurse through the widget tree to find a top level window
with the screen information (if not found check special case e.g. GtkInvisible or
screen info attached to a widget).
Rewrote GtkSelection multihead abstraction to reflect virtual atom changes and the
fact that the display information is not available a destruction time. Now it is buggy ;(
Rewrote GtkClipboard multihead abstraction to remove dependency on the display information
as it is not available at destruction time.
All the atoms are virtual in gtk, in gdk atoms are translated into virtual atoms as the
GdkEvents are constructed (the only remaining real atoms are WM related atoms and
Motif Dnd target table).
Sun May 13 12:01:12 2001 Owen Taylor <otaylor@redhat.com>
* autogen.sh (have_automake): Require libtool-1.4,
automake-1.4p1.
* acinclude.m4: Remove libtool macros.
* gdk-pixbuf/Makefile.am: Add dependencies to loade
modules.
* gdk/Makefile.am: Add dependency on libgdk_pixbuf.la
for libgdk.
* gtk/Makefile.am: Add dependency on libgdk_pixbuf.la
libgdk.la for libgtk.
* modules/input/Makefile.am: Make modules have full
dependencies.
Wed May 16 14:06:01 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtktreemodel.c (gtk_tree_path_new_from_string): fix bug
#54699 where paths weren't being checked for correctness.