Tim Janik [Tue, 3 Apr 2001 13:18:00 +0000 (13:18 +0000)]
changed gtk_item_factory_dump_items() and gtk_item_factory_dump_rc()
Tue Apr 3 13:52:57 2001 Tim Janik <timj@gtk.org>
* gtk/gtkitemfactory.[hc]: changed gtk_item_factory_dump_items()
and gtk_item_factory_dump_rc() GtkPatternSpec argument to
GPatternSpec. this is actually unlikely to cause breakage in
third-party apps since except for gle, pretty much ever caller
passes NULL here.
* gtk/gtkstyle.c (gtk_default_draw_resize_grip): assert unhandled
directions aren't passed in.
* gtk/gtksettings.[hc] (_gtk_settings_parse_convert): export conversion
functionality to be usable from gtkstyle.c as well, give precedence
for conversion to user-supplied parsers.
s/_gtk_rc_property_parser_for_type/_gtk_rc_property_parser_from_type/;
* gtk/gtkstyle.c (_gtk_style_peek_property_value): use
_gtk_settings_parse_convert() for rcporperty value conversion.
Owen Taylor [Mon, 2 Apr 2001 22:45:19 +0000 (22:45 +0000)]
Applied patch from Ron Steinke to add signal accumulators so that signals
Mon Apr 2 16:41:08 2001 Owen Taylor <otaylor@redhat.com>
* Applied patch from Ron Steinke to add signal accumulators
so that signals with the convention 'TRUE return means
handled' stop emission on a TRUE return.
* gtk/gtkmain.[ch]: Add private accumulator
_gtk_boolean_handled_accumulator, used throughout GTK+.
* gtk/gtkspinbutton.c: Add accumulator for ::output.
* gtk/gtktipsquery.c: Add accumulator for ::widget_selected
* gtk/gtkwidget.c: Add accumulators for event signals,
::drag_motion, ::drag_drop. Use _gtk_boolean_handled_accumator
for ::mnemonic-activate.
* gtk/gtkwindow.c: Add accumulator for ::frame_event
Mon Apr 2 16:24:21 2001 Owen Taylor <otaylor@redhat.com>
Mon Apr 2 16:56:15 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtkmarshal.list (VOID:BOXED,BOXED,BOXED,BOXED): new
marshaller.
* gtk/gtktreemodel.c (gtk_tree_model_range_changed): change
"changed" signal to "range_changed".
* gtk/gtktreeview.c (gtk_tree_view_set_model): fix small bug where
removing a tree from one model would remove all trees from that
model.
* gtk/gtktreeview.c (gtk_tree_view_range_changed): use range
changed signal instead.
Owen Taylor [Mon, 2 Apr 2001 03:47:25 +0000 (03:47 +0000)]
Add boxed type for GtkRequistion. Use it for ::size-request.
Sun Apr 1 21:37:22 2001 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwidget.[ch] gtk/gtktypeutils.c gtk/gtk-boxed.defs:
Add boxed type for GtkRequistion. Use it for ::size-request.
* gtk/gtkstyle.[ch] gtk/gtktypeutils.c gtk/gtk-boxed.defs:
Add a new GtkBorder structure useful for geometry properties
for widgets. Add corresponding GTK_TYPE_BORDER.
* gtk/gtkwidget.c (gtk_widget_class_install_style_property):
Support automatic parser selection like
gtk_settings_install_property_parser().
* gtk/gtksettings.c (_gtk_rc_property_select_parser): Export
functionality for use by gtk_widget_class_install_style_property.
Support GTK_TYPE_BORDER, GTK_TYPE_REQUISITION.
Sun Apr 1 20:48:59 2001 Owen Taylor <otaylor@redhat.com>
* gtk/gtkentry.c (gtk_entry_class_init): Make invisible-char
g_param_spec_unichar().
Tim Janik [Sun, 1 Apr 2001 04:05:42 +0000 (04:05 +0000)]
fixed a bunch of notifications, added reference counts around emissions
Sun Apr 1 03:28:14 2001 Tim Janik <timj@gtk.org>
* gtk/gtkwidget.c: fixed a bunch of notifications, added
reference counts around emissions where the widget is used afterwards.
added freeze/thaws around multiple properties being notified.
(_gtk_widget_get_aux_info): cleanups.
* gtk/gtksettings.c (gtk_settings_install_property): provide
default parsing functions for gdkcolor, enums and flags.
* gtk/gtkwindow.[hc]: partially reverted patch from havoc,
applied by owen.
kept a flag in GtkWindowGeometryInfo to handle user
resetting default sizes.
cleaned up ZVT comments.
bunch of assorted bug fixes, notification fixes.
(_gtk_window_reposition): make this a really internal function.
Hans Breuer [Sat, 31 Mar 2001 19:33:49 +0000 (19:33 +0000)]
disable USE_MMX for msvc build cause the assembler doesn't fit and is out
2001-03-31 Hans Breuer <hans@breuer.org>
* config.h.win32.in : disable USE_MMX for msvc build cause
the assembler doesn't fit and is out of my scope. Disable
USE_GMODULE for msvc build as wel. The right way to share
binaries on win32 would be to use libtiff.dll etc. To reduce
installation hassles IMO it's better to include all fileformats
builtin to gdk-pixbuf
* gdk-pixbuf/makefile.msc : new file
* gdk/gdk.def : updated
* gdk/win32/gdkevents-win32.c : don't erase the background if
.no_bg is set. It improves the scrolling (e.g. of testgtk main
buttons
* gdk/win32/gdkgeometry-win32.c : added comment about the above
* gdk/win32/gdkwindow-win32.c : added three new functions like
the X version. Only one is implemented, because the other two
"gdk_window_begin_(resize|move)_drag" got no docs and appear to
be default behaviour on win32 anyway ...
* gtk/gtk.def : updated
* gtk/makefile.msc.in : update for new files, use glib-genmarshal
from where it was built and add an additional rule to automagically
build gtkmarshal.[hc]
* gtk/gtkfilesel.c (open_new_dir) : don't increase n_entries
before array access for the current entry isn't finished
* gtk/gtktreeitem.c (gtk_tree_item_subtree_button_click) :
use g_return_val_if_fail
* gtk/testgtk.c (window_controls) : don't use the *wrong* size
for the static array, but let the compiler calculate it. It
makes me wonder if gcc isn't capable to catch bugs like this ...
Mike Kestner [Fri, 30 Mar 2001 03:35:47 +0000 (03:35 +0000)]
added value_changed signal which proxies for the adjustment. Cleaned
Thu Mar 29 21:20:38 2001 Mike Kestner <mkestner@ameritech.net>
* gtk/gtkspinbutton.c : added value_changed signal which proxies for the
adjustment. Cleaned redundant preconditions (spin != NULL). Added
inline API function docs.
(gtk_spin_button_new_with_range): New convenience constructor.
(gtk_spin_button_set_increments): New adjustment helper.
(gtk_spin_button_set_range): New adjustment helper.
* gtk/testgtk.c : added test of the new convenince constructor and
value_changed signals.
docs/reference/ChangeLog
2001-03-29 Mike Kestner <mkestner@ameritech.net>
* gtk/tmpl/gtkspinbutton.html: killed function docs that are now
inline.
Owen Taylor [Thu, 29 Mar 2001 23:02:30 +0000 (23:02 +0000)]
Add private function to get the aux info structure fo the widget.
Wed Mar 28 20:30:26 2001 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwidget.[ch] (_gtk_widget_get_aux_info): Add private function
to get the aux info structure fo the widget.
* gtk/gtklabel.c gtk/gtkscrolledwindow.c gtk/gtkwindow.c: Use
above function.
[ Patch from Havoc Pennington <hp@redhat.com> ]
* gdk/x11/gdkwindow-x11.c (gdk_window_get_frame_extents): function
to get the size of the window manager frame, basically the same
code that gdk_window_get_root_origin() had
(gdk_window_get_root_origin): use gdk_window_get_frame_extents()
* gtk/gtkwindow.c (gtk_window_set_default_size): use
gdk_window_resize() if the window is realized and resizeable
* gdk/x11/gdkwindow-x11.c (gdk_window_set_geometry_hints): fix
typo so that setting gravity works
* gtk/gtkwindow.c (gtk_window_move_resize): don't ever use
allocation if auto_shrink is on, even if the default size
has not changed.
Massive reordering/sorting work. Now pretty much works.
Thu Mar 29 16:29:12 2001 Jonathan Blandford <jrb@webwynk.net>
* gtk/gtkliststore.c (gtk_list_store_iter_changed): Massive
reordering/sorting work. Now pretty much works.
* gtk/gtkrbtree.c (_gtk_rbtree_reorder): Lots of work to make
reorder mostly work. Still need to get the parity correct.
* gtk/gtktreeviewcolumn.c: Listen for "model" property changed
signal, so I can listen for the sort_column_changed signal so I
can change my "sort_indicator" property. Cool.
Don't return TRUE from the event handler, that blocks all event signals on
2001-03-29 Alexander Larsson <alexl@redhat.com>
* gtk/gtktooltips.c (gtk_tooltips_event_handler):
Don't return TRUE from the event handler, that blocks all
event signals on the tooltip widget. Now you can press buttons
in toolbars again.
Owen Taylor [Mon, 26 Mar 2001 20:54:17 +0000 (20:54 +0000)]
Don't reference the widgets in the resulting list. (#11821) This makes
Mon Mar 26 15:48:45 2001 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwindow.c (gtk_window_list_toplevels): Don't
reference the widgets in the resulting list. (#11821)
This makes this consistent with gtk_container_children.
* gtk/gtkwindow.c gtk/gtkwidget.c: Fix up the two
uses of gtk_window_list_toplevels.
Havoc Pennington [Sat, 24 Mar 2001 06:41:28 +0000 (06:41 +0000)]
make motion around the ring much slower but finer-grained (gtk_hsv_focus):
2001-03-24 Havoc Pennington <hp@pobox.com>
* gtk/gtkhsv.c (gtk_hsv_move): make motion around the ring much
slower but finer-grained
(gtk_hsv_focus): get initial focus on the ring or triangle
according to tab direction
* gtk/testgtk.c (main): get rid of weird sleep(1) on exit
Havoc Pennington [Sat, 24 Mar 2001 06:10:40 +0000 (06:10 +0000)]
make this somewhat key-navigable
2001-03-23 Havoc Pennington <hp@pobox.com>
* gtk/gtkhsv.c: make this somewhat key-navigable
* gtk/gtkfontsel.c: mnemonics
* gtk/gtkgamma.c: mnemonics
* gtk/gtkcolorsel.c: add mnemonics to labels
2001-03-22 Havoc Pennington <hp@pobox.com>
Applied big patch from Seth Lytle to fix event handler return
values, slightly modified.
* gtk/gtkcalendar.c (gtk_calendar_button_press): return TRUE if
handled.
(gtk_calendar_key_press): return TRUE for GDK_space triggers
select_day
* gtk/gtkclist.c (gtk_clist_button_press): return TRUE if handled
(gtk_clist_button_release): return TRUE if handled
* gtk/gtkcolorsel.c (mouse_release):
(palette_activate):
(palette_press):
(palette_new):
(mouse_press):
(get_screen_color):
fixed return type (void to gboolean), return TRUE,
added GTK_SIGNAL_FUNC cast
* gtk/gtkcombo.c (gtk_combo_popup_button_press): fixed return type
(void to gboolean), and return TRUE
(gtk_combo_button_release): changed return value to TRUE after
gtk_grab_add on combo->popwin
* gtk/gtkctree.c (gtk_ctree_button_press): return TRUE if triggers a
collapse or expand
* gtk/gtkcurve.c (gtk_curve_graph_events): return TRUE for
button/motion
* gtk/gtkdnd.c (gtk_drag_source_event_cb): return TRUE for button
press/release and motion that trigger act
* gtk/gtktreeitem.c (gtk_tree_item_button_press): return TRUE if button
and sensitive
(gtk_tree_item_subtree_button_click): changed type to gint, and return
TRUE if sensitive
Havoc Pennington [Thu, 22 Mar 2001 20:59:37 +0000 (20:59 +0000)]
Add ability to override the focus chain for a container explicitly
2001-03-22 Havoc Pennington <hp@redhat.com>
* gtk/gtkcontainer.c (gtk_container_set_focus_chain):
(gtk_container_unset_focus_chain): Add ability to override the
focus chain for a container explicitly
Tim Janik [Thu, 22 Mar 2001 13:44:00 +0000 (13:44 +0000)]
some cleanups, fixed mnemonic_widget handling, added PROP_MNEMONIC_WIDGET.
Thu Mar 22 13:01:44 2001 Tim Janik <timj@gtk.org>
* gtk/gtklabel.[hc]: some cleanups, fixed mnemonic_widget handling,
added PROP_MNEMONIC_WIDGET.
* gtk/gtkwindow.c (gtk_window_activate_mnemonic):
(gtk_window_remove_mnemonic):
(gtk_window_add_mnemonic): fixed assertions.
(gtk_window_activate_mnemonic): constrain modifier checks to those
permitted by gtk_accelerator_get_default_mod_mask().
got rid of gtk_window_get_default_accel_group().
* gtk/gtkwidget.c (gtk_widget_real_activate_mnemonic): handle non
activatable widgets either by focussing or a warning.
added a signal accumulator to ::activate_mnemonic which stops the
emission once a handler returned TRUE.
Renamed gtk_button_new_stock() to gtk_button_new_from_stock() and removed
2001-03-21 Alexander Larsson <alexl@redhat.com>
* gtk/gtkbutton.[ch]:
* gtk/gtkdialog.c:
Renamed gtk_button_new_stock() to gtk_button_new_from_stock() and
removed accel_group argument. Renamed gtk_button_new_accel() to
gtk_button_new_with_mnemonic() and removed accel_group argument.
* gtk/gtkcheckbutton.[ch]:
New function gtk_check_button_new_with_mnemonic().
* gtk/gtkentry.c:
Override activate_mnemonic and just grab focus.
* gtk/gtkitemfactory.c:
Don't add menu uline accel group, instead use mnemonics support.
* gtk/gtklabel.[ch]:
New support for mnemonics.
* gtk/gtkmarshal.list:
Needed BOOLEAN:BOOLEAN for activate_mnemonic.
* gtk/gtkmenu.[c]:
* gtkmenushell.c:
Use mnemonics instead of accel groups for uline
support in menu items.
Removed gtk_menu_get_uline_accel_group() and
gtk_menu_ensure_uline_accel_group().
* gtk/gtkmenuitem.c:
Override activate_mnemonic to handle switching between
menu items if there are collisions.
* gtk/gtknotebook.c:
Connect to activate_mnemonic on the tab_label, so that
activating it switches to that notebook page.
* gtk/gtkwidget.[ch]:
Add activate_mnemonic signal. New function
gtk_widget_activate_mnemonic() to emit it.
Default implementation does activate/grab_focus.
* gtk/gtkwindow.[ch]:
Add support for mnemonics in windows.
New functions:
gtk_window_add_mnemonic, gtk_window_remove_mnemonic,
gtk_window_activate_mnemonic, gtk_window_set_mnemonic_modifier
Havoc Pennington [Mon, 19 Mar 2001 22:40:35 +0000 (22:40 +0000)]
Make GtkIconSource an opaque datatype, and add a bunch of accessor
2001-03-19 Havoc Pennington <hp@redhat.com>
* gtk/gtkiconfactory.c: Make GtkIconSource an opaque datatype, and
add a bunch of accessor functions. This is because we have
reasonable expectations of extending what fields it contains in
the future.
* gtk/gtkstyle.c (gtk_default_render_icon): adapt to icon source
changes
* gtk/gtkrc.c (gtk_rc_parse_icon_source): fix to use new
GtkIconSource
Tor Lillqvist [Sun, 18 Mar 2001 22:16:41 +0000 (22:16 +0000)]
New function. print_event_state, print_event): Use preprocessor macro to
2001-03-18 Tor Lillqvist <tml@iki.fi>
* gdk/win32/gdkevents-win32.c (print_window_state): New function.
print_event_state, print_event): Use preprocessor macro to make
code more compact. Print also GDK_WINDOW_STATE events.
(synthesize_crossing_events): Remove the old #ifdeffed-out code.
* gtk/gtk.def
* gtk/makefile.mingw.in: Update.
* gtk/gtkmain.h: Check for G_PLATFORM_WIN32, not only G_OS_WIN32,
to catch also Cygwin.
* gtk/gtkspinbutton.[hc]: added rc-style argument
GtkSpinButton::shadow_type.
removed spin_button->shadow_type, gtk_spin_button_set_shadow_type()
and ARG_SHADOW_TYPE as it doesn' make much sense to try to override
rc-style settings.
* gtk/gtkrc.[hc]: added gtk_rc_scanner_new() to create an rc-file
scanner with appropriate configuration, renamed GtkRcStyleClass.clone
to create_rc_style() (we don't do cloning in standard OO sense).
added per rc style properties.
* gtk/gtkstyle.[hc]: added code to retrive pspec conformant rc-style
property values and for caching those. some cleanups.
* gtk/Makefile.am: -DG_DISABLE_CONST_RETURNS.
* gtk/gtksettings.[hc]: new file for global rc-file properties
(at least currently, should get extended to support X properties
and other communication mechanisms).
* gtk/gtkwidget.[hc]: added style property support:
(gtk_widget_class_install_style_property_parser): install style
property pspec with parser function for rc-file values other
than LONG, DOUBLE or STRING.
(gtk_widget_class_install_style_property): same as above without
parser (parsers are going to be needed quite infrequently).
(gtk_widget_style_get_property): retrive style property value.
(gtk_widget_style_get_valist): same as above with varargs support,
has NOCOPY semantics.
(gtk_widget_style_get): wrapper around gtk_widget_style_get_valist().
Changed prototype to allow for getting the location of the cell relative
Thu Mar 15 18:22:44 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtkcellrenderer.h: Changed prototype to allow for getting
the location of the cell relative to its area.
* gtk/gtkcell*: modified for above change
* gtk/gtktreeview.c: modified for above change.
Wed Mar 14 13:58:32 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtktreeview.c (gtk_tree_view_set_model): Ref the model.
(gtk_tree_view_finalize): actually unref the model. Thanks to
Jamie Strachan <frostfreek@yahoo.com> for noticing this error.
Mon Mar 12 21:02:08 2001 Jonathan Blandford <jrb@redhat.com>
* gtk/gtktreeview.c (gtk_tree_view_header_focus): Fix up focus
code a lot.
* gtk/gtktreedatalist.c (_gtk_tree_data_list_free): use
g_mem_chunk.
(_gtk_tree_data_list_alloc): use g_mem_chunk.
(_gtk_tree_data_list_node_to_value): Fix to switch on
G_TYPE_FUNDAMENTAL.
(_gtk_tree_data_list_value_to_node): Fix to switch on
G_TYPE_FUNDAMENTAL.
(_gtk_tree_data_list_node_copy): Fix to switch on
G_TYPE_FUNDAMENTAL.
Havoc Pennington [Mon, 12 Mar 2001 21:01:53 +0000 (21:01 +0000)]
Patch from John Margaglione to convert to new property API. Also, change
2001-03-12 Havoc Pennington <hp@redhat.com>
* gtk/gtkwindow.c (gtk_window_class_init): Patch from John Margaglione
to convert to new property API. Also, change default_width and
default_height get_property to return 0 as the flag value for
"unset"; -1 is not a valid value for these fields.
* gdk/linux-fb/gdkdrawable-fb2.c:
window->mapped to GDK_WINDOW_IS_MAPPED().
* gdk/linux-fb/gdkwindow-fb.c:
window->mapped to GDK_WINDOW_IS_MAPPED().
Added gdk_window_shape_combine_region.
Added assorted placeholders for unimplemented new GdkWindow API calls.
Add more test when converting filenames from/to utf-8. Don't allow
2001-03-12 Alexander Larsson <alexl@redhat.com>
* gtk/gtkfilesel.c:
Add more test when converting filenames from/to utf-8.
Don't allow entering text in the filename entry which isn't allowed in
a filename. Just beep instead.
Hans Breuer [Sat, 10 Mar 2001 18:13:03 +0000 (18:13 +0000)]
updated
2001-03-10 Hans Breuer <hans@breuer.org>
* gdk/gdk.def : updated
* gdk/win32/gdkgc-win32.c : pixmaps are visible again, due
to offseting the clip_mask at the right place
* gdk/win32/gdkwindow-win32.c : don't create background brushes
for all the gdkClasses, especially not for the input_only case.
Otherwise flickering wont be avoidable, because we don't get
WM_ERASEBKGND message for these windows.
* gdk/win32/gdkevents-win32.c : initialize event->expose.region
and fixed massive redraw problems apparently triggered by Alexanders
recent change. Not fully understood the GDK_EXPOSE / WM_PAINT handling
but it seems to work quite well :-)
* gdk/win32/gdkwindow-win32.c : reflect GdkWindowObject::mapped
removing and implemented Havoc's GdkWindow state functions;
stick/unstick only as no-op, cause I don't know how to emulate
on win32 ...
* gtk/gtk.def : updated
* gtk/gtkbox.c (gtk_box_get_spacing) : use g_return_val_if_fail ()
Document expose event->region change and that gtk_widget_event doesn't
2001-03-09 Alexander Larsson <alexl@redhat.com>
* docs/Changes-2.0.txt:
Document expose event->region change and that gtk_widget_event
doesn't allow passing expose events.
* gdk/gdkevents.h:
Add region to expose event.
* gdk/gdkevents.c:
Handle event->region in gdk_event_copy() and gdk_event_free().
* gdk/gdkwindow.c (gdk_window_process_updates_internal):
Generate expose_event->region when creating expose events.
* gdk/x11/gdkevents-x11.c:
Generate expose events with regions when translating X events.
* gtk/gtkcontainer.[ch]:
Default expose handler that propagates expose events to NO_WINDOW
children. New function gtk_container_propagate_expose()
* gtk/gtkwidget.[ch]:
Moved gtk_widget_event implementation to gtk_widget_event_internal.
gtk_widget_event calls gtk_widget_event_internal but doesn't allow
expose events. New function gtk_widget_send_expose() to send expose
events. New function gtk_widget_region_intersect() to calculate
window/region intersections.
* gtk/gtkmain.c (gtk_main_do_event):
Use gdk_window_begin_paint_region() instead of gdk_window_begin_paint_rect().
Use gtk_widget_send_expose() to send expose events.
projects / ~andy / gtk / log