Owen Taylor [Thu, 16 May 2002 23:59:23 +0000 (23:59 +0000)]
Only optimize out the full configure handling if we got a gratuitous
Thu May 16 19:49:24 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwindow.c (gtk_window_configure_event): Only
optimize out the full configure handling if we got
a gratuitous configure event(configure_request_count == 0)
and the size didn't change.
* gtk/gtkwindow.c (gtk_window_move_resize): Use
g_message(), not g_print() for debug spew.b
* gtk/gtkwindow.c (gtk_window_real_set_focus): Ref
the old and new focus widgets for the scope of
the function.
* gtk/gtksocket.c gtkplug.c: Remove various #if 0'd
bits of old code.
Owen Taylor [Thu, 16 May 2002 23:38:01 +0000 (23:38 +0000)]
Thu May 16 18:47:29 2002 Owen Taylor <otaylor@redhat.com> * gtk/gtkwindow.c (gtk_window_show): Disable the focus-some-widget code for GtkPlug descendants. * gtk/gtkplug.c (gtk_plug_focus_event): Eat focus-in/out events since they can be generated by keyboard grabs, but are spurious for an embeddded widget. (#79196, Padraig O'Briain) * gtk/gtksocket.c (gtk_socket_size_request): Only pay attention to the minimum size hints, and changes to that ... ignore the default size. * gtk/gtksocket.c (gtk_socket_send_configure_event): Pass in root-window relative coordinates as per the ICCCM. * gtk/gtksocket.c (gtk_socket_size_allocate): Send as many synthetic configures as necessary to make up for ConfigureRequest events we got. (#78799, Dave Camp.)
Thu May 16 23:59:56 2002 Kristian Rietveld <kris@gtk.org>
Merged from gtk-2-0.
* gtk/gtktreestore.c (gtk_tree_store_sort_iter_changed): set the prev
of the next node correctly, so we don't corrupt the list and cause
bad things. (fixes #71374)
Owen Taylor [Thu, 16 May 2002 16:54:03 +0000 (16:54 +0000)]
gdk_event_put() copies the event and setting name, so we don't have to
Thu May 16 12:50:13 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/gdkevents-x11.c (gdk_xsettings_notify_cb):
gdk_event_put() copies the event and setting name,
so we don't have to strdup() it here. (Fix from
Michael Meeks, #81953)
Ettore Perazzoli [Thu, 16 May 2002 14:00:45 +0000 (14:00 +0000)]
If the menu is scrolled down or up, and some of the widget area is empty,
* gtk/gtkmenu.c (gtk_menu_paint): If the menu is scrolled down or
up, and some of the widget area is empty, don't fill it in with
GTK_SHADOW_IN rectangles that make it looks strange. Fixes
#72695.
Owen Taylor [Wed, 15 May 2002 22:27:42 +0000 (22:27 +0000)]
Don't call gdk_pointer_grab() to change the cursor if we've already
Wed May 15 18:15:45 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkdnd.c (_gtk_drag_source_handle_event): Don't
call gdk_pointer_grab() to change the cursor if we've
already ungrabbed on button release / escape.
(#80420, Dave Camp)
Owen Taylor [Tue, 14 May 2002 23:57:40 +0000 (23:57 +0000)]
Subsitute in @PERL@ instead of hardcoded /usr/bin/perl. (#78355, Gareth
Tue May 14 19:53:15 2002 Owen Taylor <otaylor@redhat.com>
* demos/gtk-demo/geninclude.pl.in configure.in:
Subsitute in @PERL@ instead of hardcoded
/usr/bin/perl. (#78355, Gareth Pearce. Fix based
on patch from Matthias Clasen)
* gdk/Makefile.am: Patch from Daniel Elstner to use glib-mkenums
instead of makeenums.pl.
* gdk/{x11,linux-fb,win32}/Makefile.am: Hack to
get gdkenumtypes.h built first.
Owen Taylor [Tue, 14 May 2002 22:44:51 +0000 (22:44 +0000)]
Remove the result file on error. (#79929, David Highley)
Tue May 14 18:41:58 2002 Owen Taylor <otaylor@redhat.com>
* gtk/stock-icons/Makefile.am (gtkstockpixbufs.h):
Remove the result file on error. (#79929, David
Highley)
* gtk/gtkoptionmenu.c (gtk_option_menu_remove_menu):
If the menu is currently active, cancel it before
setting the new menu. (#79162, Jonathan Blandford.)
Owen Taylor [Tue, 14 May 2002 21:36:32 +0000 (21:36 +0000)]
Fix format of selection data to be 8 rather than GTK_TYPE_POINTER.
Tue May 14 17:30:57 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkctree.c gtk/gtkclist.c: Fix format of
selection data to be 8 rather than GTK_TYPE_POINTER.
(#79751, James Henstridge)
* gtk/gtkmenu.c (gtk_menu_key_press): Call
gtk_accelerator_valid() with cleaned up
accel_key/accel_mods, not the raw values from
the event. (#80513, Havoc Pennington.)
* gtk/gtkmenu.c (gtk_menu_set_tearoff_state): Remove
some dead code that set the tearoff window not
resizable and removed decorations. (#80235)
Owen Taylor [Tue, 14 May 2002 20:55:22 +0000 (20:55 +0000)]
Set the width of the layout to the actual wrap width (our requisition) not
Tue May 14 16:44:09 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtklabel.c (gtk_label_ensure_layout): Set
the width of the layout to the actual wrap width
(our requisition) not the width we set when calculating
the width. This results in the lines being justified
within the correct area. (#79157, Anders Carlsson)
* gtk/gtkaccelgroup.c gtk/gtkmarshelers.list: Corrected
registered parameter types of "accel_activate" and
"accel_changed" signals. (Patch from Murray Cumming, #78798)
* gtk/gtkrc.c (gtk_rc_make_default_dir): Switch
binary-version and type to correspond to the
current ordering in _gtk_get_module_path.
(#78746, Sergey Kuzminov)
* gtk/gtkrc.c (gtk_rc_get_module_dir): Add docs
pointing to the GTK_PATH documentation.
* gtk/Makefile.am (uninstall-local): Delete
key themes as well. (#81286, Kristian Rietveld.)
Owen Taylor [Tue, 14 May 2002 19:45:08 +0000 (19:45 +0000)]
Add a sanity g_return_if_fail() check for the depths for the window and
Tue May 14 15:40:49 2002 Owen Taylor <otaylor@redhat.com>
* gdk/x11/gdkwindow-{linux-fb,win32,x11.c} (gdk_window_set_back_pixmap): Add a sanity g_return_if_fail() check for the depths
for the window and pixmap. (#79191, Shivaram Upadhyayula)
Matthias Clasen [Mon, 13 May 2002 22:35:42 +0000 (22:35 +0000)]
Change parameter names to match the header file and the documentation.
* gtk/gtkcontainer.c (gtk_container_class_install_child_property):
(gtk_container_class_find_child_property):
(gtk_container_class_list_child_properties):
* gtk/gtkwidget.c (gtk_widget_class_find_style_property):
(gtk_widget_class_list_style_properties): Change parameter names
to match the header file and the documentation.
Daniel Elstner [Fri, 10 May 2002 21:59:20 +0000 (21:59 +0000)]
[ merge from stable ]
2002-05-10 Daniel Elstner <daniel.elstner@gmx.net>
[ merge from stable ]
* gtk/gtktreemodel.c: #define INITIALIZE_TREE_ITER(), which zeroes out all GtkTreeIter fields. (#81380)
(gtk_tree_model_get_iter): Use it.
(gtk_tree_model_iter_children): ditto
(gtk_tree_model_iter_nth_child): ditto
(gtk_tree_model_iter_parent): ditto
Wed May 8 16:14:49 2002 Soeren Sandmann <sandmann@daimi.au.dk>
* gtk/gtktreeview.c
(gtk_tree_view_init): set_redraw_on_allocate (FALSE);
(gtk_tree_view_size_allocate): invalidate last column if width
changed.
(gtk_tree_view_bin_expose): skip over cells that aren't in
event->region
added a new conditional CROSS_COMPILING which indicates ($build != $host).
2002-05-07 Michael Natterer <mitch@gimp.org>
* configure.in: added a new conditional CROSS_COMPILING which
indicates ($build != $host). If it is set, look for
gdk-pixbuf-csource in PATH. Error out if it was not found.
* gtk/stock-icons/Makefile.am: use the gdk-pixbuf-csource found at
configure time if CROSS_COMPILING is set, use the one which was
just built otherwise.
build the static pixbuf loaders conditionally like the dynamic ones so
2002-05-07 Michael Natterer <mitch@gimp.org>
* gdk-pixbuf/Makefile.am: build the static pixbuf loaders
conditionally like the dynamic ones so --without-lib<foo> is
honored when building with --disable-modules. (#66867)
* gtk/gtktreeview.c (gtk_tree_view_button_press): focus on a cell
if applicable,
(gtk_tree_view_bin_expose): set_cell_data before iterating columns,
add support for row-spanning focus rectangles,
(gtk_tree_view_has_special_cell): new function,
(gtk_tree_view_move_cursor_left_right): add support for multiple
focusable cells in one column,
(gtk_tree_view_set_cursor): call _set_cursor_on_cell now,
(gtk_tree_view_set_cursor_on_cell): copy of _set_cursor, extended
with focus_cell parameter,
(gtk_tree_view_search_iter): removed unused column variable,
(gtk_tree_view_start_editing): add neighbor size code to allow
for multiple editable cells in one column.
* gtk/gtktreeviewcolumn.c (_GtkTreeViewColumnCellInfo): add
in_editing_mode field,
(gtk_tree_view_column_get_edited_cell): new function, removed
_get_editable_cell,
(_gtk_tree_view_column_get_cell_at_pos): new function,
(gtk_tree_view_column_pack_end): s/g_new/g_new0/ ...,
(gtk_tree_view_column_cell_process_action): loads of changes to get
it right and to allow for multiple special cells, etc,
(gtk_tree_view_column_cell_first): new function,
(gtk_tree_view_column_cell_last): ditto,
(gtk_tree_view_column_cell_next): ditto,
(gtk_tree_view_column_cell_prev): ditto,
(gtk_tree_view_column_cell_focus): add left and right parameters,
allow for multiple special cells,
(gtk_tree_view_column_cell_is_visible): add assertion,
(gtk_tree_view_column_focus_cell): new function,
(gtk_tree_view_column_stop_editing): unset in_editing_mode flag,
(_gtk_tree_view_column_get_neighbor_sizes): iterate through cells
correctly
* gtk/gtktreeprivate.h: s/_get_editable_cell/_get_edited_cell/, add
_gtk_tree_view_column_get_cell_at_pos, add new parameters to
_gtk_tree_view_column_cell_focus.
* tests/testtreeedit.c: add some cells in order to test new code.
Soeren Sandmann [Sun, 5 May 2002 00:59:42 +0000 (00:59 +0000)]
image files for testing pixbuf loaders the old test-loaders.c split into
Sun May 5 02:22:59 2002 Soeren Sandmann <sandmann@daimi.au.dk>
* tests/test-images/*: image files for testing pixbuf loaders
* tests/pixbuf-*.c: the old test-loaders.c split into four
different programs
* tests/Makefile.am: add new pixbuf tests
Sun May 5 02:27:01 2002 Soeren Sandmann <sandmann@daimi.au.dk>
* gdk/x11/gdkvisual-x11.c:
* gdk/x11/gdkwindow-x11.c:
* gdk/x11/gdkdisplay-x11.[hc] (gdk_open_display) :
Rationalized the screen initialization process,
changed the GdkDisplayX11 screen list to an array.
* gdk/x11/gdkscreen-x11.[hc] (_gdk_x11_screen_new) :
Moved the xinerama init functions to gdkdisplay-x11.c to
this file, create a single function to fully initialize
a GdkScreen. #79981
* gdk/gdkscreen.[hc] (gdk_screen_x11_get_monitor_at_window):
remove virtualization and made it cross platform. fix #79961
* gdk/x11/gdkdisplay-x11.c (check_solaris_xinerama):
fixed initialization bugs.
* gdk/x11/gdkscreen-x11.[hc] (gdk_screen_get_monitor_geometry):
* tests/testxinerama.c:
* gtk/gtkwindow.c (gtk_window_compute_configure_request):
* gtk/gtkmenu.c (gtk_menu_position):
take a GdkRectangle argument instead of returning
a GdkRectangle * and propagated the changes. fix #79974
Owen Taylor [Tue, 30 Apr 2002 23:44:14 +0000 (23:44 +0000)]
Make it work, even if the window is already realized or mapped.
Tue Apr 30 18:42:25 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwindow.c (gtk_window_set_screen): Make it
work, even if the window is already realized or
mapped.
* tests/testgtk.c: Erwann's multihead changes, slightly
dumbed down. (Removed logic for pulling windows onto
the current screen, people can click twice on the demo
to destroy and create again.)
* tests/prop-editor.c: Set the screen on the property
editor if the edited object is a widget.
Owen Taylor [Tue, 30 Apr 2002 23:43:58 +0000 (23:43 +0000)]
Make it work, even if the window is already realized or mapped.
Tue Apr 30 18:42:25 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwindow.c (gtk_window_set_screen): Make it
work, even if the window is already realized or
mapped.
* tests/testgtk.c: Erwann's multihead changes, slightly
dumbed down. (Removed logic for pulling windows onto
the current screen, people can click twice on the demo
to destroy and create again.)
* tests/prop-editor.c: Set the screen on the property
editor if the edited object is a widget.
use gtk_text_layout_invalidate_cache instead of cut-and-paste
2002-04-30 Havoc Pennington <hp@redhat.com>
* gtk/gtktextlayout.c (gtk_text_layout_real_free_line_data): use
gtk_text_layout_invalidate_cache instead of cut-and-paste
(gtk_text_layout_real_invalidate): always invalidate cache, even
if the line has no stored line data. #78573
Owen Taylor [Tue, 30 Apr 2002 20:17:36 +0000 (20:17 +0000)]
Multi-screen test case form multihead branch with a few cleanups. (Get rid
Tue Apr 30 15:59:53 2002 Owen Taylor <otaylor@redhat.com>
* tests/testmultiscreen.c: Multi-screen test case
form multihead branch with a few cleanups. (Get rid
of use of g_object_connect())
* tests/testmultidisplay.c: Multi-display test case
from multihead branch with a few improvements.
(Allow the other display name to be passed on
the command line, use a GtkDialog to get more attractive
spacing.)
half-finished docs on merging GTK with existing hunks of code (event loop
2002-04-30 Havoc Pennington <hp@redhat.com>
* gtk/other_software.sgml: half-finished docs on merging GTK with
existing hunks of code (event loop integration, using threads
vs. processes, blah blah)
assign something to "prev" so that removing tag info succeeds. Part of
2002-04-30 Havoc Pennington <hp@redhat.com>
* gtk/gtktextbtree.c (gtk_text_btree_remove_tag_info): assign
something to "prev" so that removing tag info succeeds.
Part of #77301
* gtk/gtktextbtree.c (_gtk_text_btree_unref): reorder so that it
unrefs the tag table first, so that the btree is in a consistent
state when we're removing tags from it.
* gtk/gtktexttagtable.c (_gtk_text_tag_table_remove_buffer): strip
all tags in the tag table out of the buffer. #77301
Alex Larsson [Mon, 29 Apr 2002 22:53:45 +0000 (22:53 +0000)]
Integrate Erwann Chenede's multihead changes for the gtk/ directory.
Mon Apr 29 18:28:00 2002 Owen Taylor <otaylor@redhat.com>
Integrate Erwann Chenede's multihead changes for the gtk/ directory.
* gtk/gtkclipboard.[ch]: Add gtk_clipboard_get_for_display(),
make internals multihead aware.
* gtk/gtkcolorsel.[ch]: Add
gtk_color_selection_set_change_palette_with_screen_hook () [ugh!]
make up for non-multihead safety of
gtk_color_selection_set_change_palette_hook()
* gtk/gtkinvisible.[ch] gtk/gtkmenu.[ch] gtkwindow.[ch]: Add
gtk_{invisible,menu,window}_set_screen(); add "screen" properties
for GtkWindow and GtkMenu.
* gtk/gtkclipboard.c: Use a GtkImage rather than a pixmap for
the dropper, look up the color palette only at realization time,
other multihead fixes.
* gtk/gtkcombo.c (gtk_combo_unrealize): Popdown the list when
unrealizing.
* gtk/gtkentry.c: Only claim ownership of the primary selection
when realized, misc multihead fixes.
* gtk/gtkfontsel.c: Only fill in fonts when attached to a screen,
fix gtk_font_selection_get_font() for multihead.
* gtk/gtkgc.c: make the depth => drawable hash per-screen.
* gtk/gtkinvisible.c: Add a constructor that realizes the
widget, so we get a realized widget with g_object_new() as
well gtk_invisible_new() as before.
* gtk/gtkmain.c: Get rid of unused gtk_visual/gtk_colormap
variables.
* gtk/gtktextdisplay.c: Add warnings if stipple bitmaps
are used on the wrong screen.
* gtk/gtktoolbar.c: Make handling of GtkSettings-based layout
read properties and connect to settings when the screen is changed,
rather than on init/finalize.
* gtk/gtkwindow.c: Fix icon handing to be multihead safe ...
default icon pixmaps/mask are only shared between windows on the
same screen. Misc multihead fixes.
Sat Apr 27 13:49:53 2002 Owen Taylor <otaylor@redhat.com>
* gtk/gtkclipboard.c (gtk_clipboard_get_for_display):
Update docs to reference GDK_SELECTION_CLIPBOARD rather GDK_NONE.
2002-04-29 Alex Larsson <alexl@redhat.com>
* gdk/linux-fb/gdkproperty-fb.c (gdk_property_get):
Fix silly bug, noticed by Sven Neumann.
Sun Apr 28 22:43:55 2002 Jonathan Blandford <jrb@gnome.org>
* gtk/gtktreemodelsort.c (gtk_tree_model_sort_set_sort_func): Fix
so that you can set a new sort func.
Anders Carlsson [Fri, 26 Apr 2002 22:41:17 +0000 (22:41 +0000)]
Use GDK_DISPLAY_XDISPLAY instead of GDK_SCREEN_XDISPLAY since we pass in a
2002-04-27 Anders Carlsson <andersca@gnu.org>
* gdk/x11/gdkpixmap-x11.c (gdk_pixmap_foreign_new_for_display):
Use GDK_DISPLAY_XDISPLAY instead of GDK_SCREEN_XDISPLAY since
we pass in a GdkDisplay * in the macro.
* gdk/gdkinternals.h:
* gdk/x11/gdkmain-x11.c: (_gdk_windowing_set_default_display):
New function that sets gdk_display to the default display.
Fixes #79965
if the deleted node was "below" this path, then do not unref all nodes
Fri Apr 26 21:58:16 2002 Kristian Rietveld <kris@gtk.org>
* gtk/gtktreemodel.c (gtk_tree_row_ref_deleted_callback): if the
deleted node was "below" this path, then do not unref all nodes
between the deleted parent and the node this rowref is pointing too
(see added comment in gtk_tree_row_ref_deleted_callback),
(fixes #77977)
Sven Neumann [Fri, 26 Apr 2002 19:50:17 +0000 (19:50 +0000)]
tests/testtreecolumns.c quit the main loop when the window is destroyed.
2002-04-26 Sven Neumann <sven@gimp.org>
* tests/testtreecolumns.c
* tests/testtreeview.c: quit the main loop when the window is
destroyed. Removed compiler warnings about wrong format in printf().
This commit fixes binary compatibility, which I broke with my recent cell
Fri Apr 26 21:46:42 2002 Kristian Rietveld <kris@gtk.org>
This commit fixes binary compatibility, which I broke with my
recent cell renderer commits :/.
* gtk/gtkcellrenderer.h: move cell_background out of _GtkCellRenderer
* gtk/gtkcellrenderer.c: create GtkCellRendererInfo, with the
cell_background field and use this as object data. This way we keep
the field private and dont break ABI.
* gtk/gtkcellrendererpixbuf.h: remove stock_id, stock_size and
stock_detail from _GtkCellRendererPixbuf
* gtk/gtkcellrendererpixbuf.c: create a GtkCellRendererPixbufInfo
with above mentioned fields. And update everything.
Owen Taylor [Thu, 25 Apr 2002 22:29:14 +0000 (22:29 +0000)]
Start of integration of Erwann Chenede's multihead work from the
Thu Apr 25 16:51:40 2002 Owen Taylor <otaylor@redhat.com>
Start of integration of Erwann Chenede's multihead work
from the gtk-multihead branch.
* gdk/gdkdisplay.[ch] gdk/gdkscreen.[ch]
gdk/x11/gdkdisplay-x11.[ch] gdk/x11/gdkscreen-x11.[ch]
New classes representing a set of screens with attached
input devices and a single contiguous area, respectively.
* gdk/gdk.[ch] gdk/gdkinternals.h gdk/x11/gdkmain-x11.c:
gdk/x11/gdkprivate-x11.h: Make the initialization interface
simple _gdk_windowing_init() and do the rest in
gdk_open_display() calls.
* gdk/gdk.[ch]: Add gdk_parse_args() which can be used
to do the display-independent part of initialization
instead of gdk_init_[check].
* gdk/gdkcursor.h gdk/gdkfont.h gdk/gdkkeys.h gdk/gdkpixmap.h
gdk/gdkproperty.h gdk/gdkselection.h gdk/gdkwindow.h:
Add multihead variants (_for_display(), for_screen()) of functions
getting information specific to a particular screen screen or
display.
* gdk/gdkscreen.[ch]: Add gdk_screen__* variants of functions
like gdk_rgb_get_colormap() that used to get/list global
objects.
* gdk/x11/gdkx.h: Add functions for converting GdkScreen
and GdkDisplay into the X equivalents.
* gdk/x11/gdkwindow-x11.c: Removed gdk_window_xid_at_coords()
not in the headers and unused.
* configure.in gdk/x11/{gxid.c,gxid_lib.[ch],gdkinput-gxi.c}:
Remove gxid support ... has not been tested for a long time...
"xfree" support is more portable to non XFree86.
* gdk/**.h: Add a GDK_MULTIHEAD_SAFE define that can be
used to turn off functions that are inherently non-multihead
safe.
* gdk/**.c: add GDK_NOTE(multihead, ...) calls when functions
are used in non-multihead-safe ways.
* gdk/*.c gdk/x11/*.c: Changes to make the internals of GDK
multihead safe.