Tim Janik [Mon, 27 Jul 1998 01:50:16 +0000 (01:50 +0000)]
fixup, so that enum values are looked up.
Mon Jul 27 03:11:20 1998 Tim Janik <timj@gtk.org>
* gtk/gtkbindings.c (binding_compose_params): fixup, so that enum
values are looked up.
* gtk/gtktypeutils.h:
* gtk/gtktypeutils.c: new functions gtk_type_flags_find_value and
gtk_type_enums_find_value to lookup enum values.
* gtk/gtkctree.h:
* gtk/gtkctree.c: change enum name from GtkCTreeExpansion to
GtkCTreeExpansionType. fixed signal creations to pass the enum
types not just the base type GTK_TYPE_ENUM.
* gtk/gtkclist.h:
* gtk/gtkclist.c: fixed signal creations to pass the real enum
types not just the base type GTK_TYPE_ENUM.
Stefan Jeske [Sun, 26 Jul 1998 14:45:40 +0000 (14:45 +0000)]
Began cleanup, e.g. use gtk_widget_show_all instead of many
gtk_widget_show's. Completely rewrote the button box sample
(only one window now). Removed "miscellaneous" button.
Tim Janik [Wed, 22 Jul 1998 22:29:10 +0000 (22:29 +0000)]
introduce testbindings on C-1 for buttons, to check out binding
Thu Jul 23 00:11:39 1998 Tim Janik <timj@gtk.org>
* gtk/testgtkrc: introduce testbindings on C-1 for buttons,
to check out binding priorities. someone should really write
gtkrc-mode for emacs.
* gtk/gtkwidget.c (gtk_widget_class_init): remove ugly bindings test.
* gtk/testgtk.c (main): and move it here ;). this test says something
on release of C-9. (this works only if the mouse pointer is on a
spinbutton and there is no focus widget).
* gtk/gtkrc.h:
* gtk/gtkrc.c:
export gtk_rc_parse_color, gtk_rc_parse_state and a new function
gtk_rc_parse_priority to parse path priority types.
export rc tokens.
feature binding parsing.
Wed Jul 22 23:41:17 1998 Tim Janik <timj@gtk.org>
* gtk/gtkbindings.h:
* gtk/gtkbindings.c: new function gtk_binding_parse_binding() to
feature binding parsing in rc files.
* gtk/gtkbindings.c (gtk_binding_set_new): bug fix.
* gtk/gtkbindings.c (gtk_binding_pattern_compare): bug fix, we used to
order the bindings with lowest priority first.
actually builds and runs properly on big endian machines now, because we
* gdk/gdkrgb.c: actually builds and runs properly on big endian machines
now, because we #include config.h. Also fixed up a compile error in the
WORDS_BIGENDIAN case in gdk_rgb_convert_565_gray. I made the fix by looking
at patterns in the other code; Raph can you verify this?
* gdk/gdkthreads.c: #include <unistd.h> to quell compiler warnings
Elliot Lee [Tue, 21 Jul 1998 18:49:53 +0000 (18:49 +0000)]
genmarshal.pl: Handle complex types, and functions with a return value but
genmarshal.pl: Handle complex types, and functions with a return value but
no parameters.
genmarshal.pl gtkmarshal.* gtkwidget.c: Remove support for GDK_EVENT and
STYLE parameter types.
Stefan Jeske [Tue, 21 Jul 1998 04:49:01 +0000 (04:49 +0000)]
Fixed wrong size of activity indicator. It is settable now as a fraction
Tue Jul 21 06:32:43 1998 Stefan Jeske <stefan@gtk.org>
* gtk/gtkprogressbar.{c,h} gtk/testgtk.c: Fixed wrong size of
activity indicator. It is settable now as a fraction of the
total width/height using gtk_progress_bar_set_activity_blocks.
Renamed gtk_progress_bar_set_number_of_blocks to
gtk_progress_bar_set_discrete_blocks.
Elliot Lee [Tue, 21 Jul 1998 04:13:42 +0000 (04:13 +0000)]
Use common marshalling routines instead of having widget-private ones. It
Use common marshalling routines instead of having widget-private ones.
It compiles & links.
testgtk segfaults. I'm convinced that is Somebody Else's Problem
(marshalling routine gets func_data=0x0 and blithely passes it on) but the
fact that it happens now & not before makes me wrong :-)
Lars Hamann [Mon, 20 Jul 1998 19:45:45 +0000 (19:45 +0000)]
new functions, recursive process tree to specified depth
* gtk/gtkctree.c (gtk_ctree_post_recursive_to_depth)
(gtk_ctree_pre_recursive_to_depth): new functions, recursive process
tree to specified depth
(gtk_ctree_expand_to_depth): expand tree to specified depth
(gtk_ctree_collapse_to_depth): collapse all nodes with level >= depth
Tim Janik [Sun, 19 Jul 1998 19:09:27 +0000 (19:09 +0000)]
export gtk_binary_age and gtk_interface_age.
Sat Jul 18 22:59:04 1998 Tim Janik <timj@gtk.org>
* gtk/gtkmain.h:
* gtk/gtkmain.c: export gtk_binary_age and gtk_interface_age.
* gtk/gtktypeutils.h:
* gtk/gtkitemfactory.h: remove "void" from prototypes again.
Sat Jul 18 00:54:01 1998 Lars Hamann <lars@gtk.org>
* gtk/gtknotebook.c (gtk_notebook_set_child_arg): added support for
CHILD_ARG_MENU_LABEL, CHILD_ARG_TAB_LABEL
(gtk_notebook_get_child_arg): added support for CHILD_ARG_MENU_LABEL,
CHILD_ARG_POSITION
(gtk_notebook_reorder_child): new funktion to change position of a
notebook page
(gtk_notebook_draw_arrow): draw left/right arrow with
GTK_SHADOW_ETCHED_IN if focus_tab is first/last tab.
Stefan Jeske [Sun, 19 Jul 1998 10:35:39 +0000 (10:35 +0000)]
Introduced generic GtkProgress widget. Derived GtkProgressBar from
Sun Jul 19 12:19:16 1998 Stefan Jeske <stefan@gtk.org>
* gtk/gtkprogress.c gtk/gtkprogress.h gtk/gtkprogressbar.c
gtk/gtkprogressbar.h gtk/Makefile.am gtk/gtk.h gtk/testgtk.c:
Introduced generic GtkProgress widget. Derived GtkProgressBar from
GtkProgress. Made GtkProgressBar much more flexible (see testgtk).
Tim Janik [Fri, 17 Jul 1998 06:10:02 +0000 (06:10 +0000)]
sropped the approach of copying all signal specific data into a
Fri Jul 17 05:30:47 1998 Tim Janik <timj@gtk.org>
* gtk/gtksignal.c (gtk_handlers_run) (gtk_signal_real_emit):
sropped the approach of copying all signal specific data into a
GtkHandlerInfo structure. we rather keep the signal data on the stack
in gtk_signal_real_emit and pass that over to gtk_handlers_run. this
avoids multiple lookups of the signal pointer and works savely together
with destroy notifiers of handlers.
* gtk/gtkentry.c (gtk_entry_key_press): care for the posssibility that
someone destroys the entry on the ::activate signal.
Tim Janik [Fri, 17 Jul 1998 00:03:35 +0000 (00:03 +0000)]
use misc->xpad for x calculation instead of ypad (typo). (gtk_arrow_set):
Fri Jul 17 01:21:32 1998 Tim Janik <timj@gtk.org>
* gtk/gtkarrow.c (gtk_arrow_expose): use misc->xpad for x calculation
instead of ypad (typo).
(gtk_arrow_set): clear all of the arrow's allocation (left a 1 pixel
border around the allocation).
(gtk_arrow_init): the static requisition needs to be setup in this
function rather than gtk_widget_new(), so arrow creation works without
the gtk_arrow_new() constructing.
(gtk_arrow_expose): return TRUE, since we actually handled the expose.
* gtk/gtkwidget.c (gtk_widget_real_size_request): provide a default
implementation for GtkWidget::size_request, so size requisition for
simple statically sized widgets like GtkArrow works.
minor changes to gtk-config: --cflags doesn't prepend -I@libdir@/glib/include
anymore, since glib-config handles it; --libs strips out redundant -L<path>'s
properly (there was code in there to do it before, but it didn't work)
Tim Janik [Thu, 16 Jul 1998 20:23:12 +0000 (20:23 +0000)]
signdness corrections all ove the place. implementation of object
Thu Jul 16 18:24:41 1998 Tim Janik <timj@gtk.org>
* gtk/gtknotebook.h:
* gtk/gtknotebook.c: signdness corrections all ove the place.
implementation of object arguments: GtkNotebook::enable_popup,
GtkNotebook::tab_pos, GtkNotebook::tab_border, GtkNotebook::show_tabs,
GtkNotebook::show_border, GtkNotebook::scrollable and GtkNotebook::page.
begun implementation of child args, we have GtkNotebook::tab_label, but
GtkNotebook::menu_label and GtkNotebook::position still lack the
implementation.
(gtk_notebook_remove_page):
(gtk_notebook_current_page): allow negative indices (-1) as alias for
the last page.
* gtk/gtkentry.h:
* gtk/gtkentry.c: GtkType and macro fixups. implementation of object
arguments GtkEntry::max_length and GtkEntry::visibility.
(gtk_entry_adjust_scroll): queue a redraw.
* gtk/gtkeditable.h:
* gtk/gtkeditable.c: GtkType und macro fixups. implementation of
object arguments GtkEditable::text_position and GtkEditable::editable.
new (action) signal GtkEditable::set_editable.
provide a default handler for GtkEditable::changed that queues a redraw.
Lars Hamann [Thu, 16 Jul 1998 13:05:02 +0000 (13:05 +0000)]
fix for sigsegv in case of drag_target == NULL fix for
* gtk/gtkctree.c (gtk_ctree_button_release): fix for sigsegv in
case of drag_target == NULL
* gtk/gtkclist.c (gtk_clist_button_press): fix for GTK_2/3BUTTON_PRESS.
Clist did not emit an select_row signal in case of GTK_2/3BUTTON_PRESS.
* gtk/gtkfilesel.c (gtk_file_selection_file_button,
gtk_dir_selection_file_button): fix due to changes in clist's
select_row signal.
Tim Janik [Thu, 16 Jul 1998 02:47:15 +0000 (02:47 +0000)]
handle negative values similar to gtk_widget_set_uposition(). that is:
Thu Jul 16 01:27:15 1998 Tim Janik <timj@gtk.org>
* gtk/gtkwidget.h:
* gtk/gtkwidget.c (gtk_widget_set_usize): handle negative values
similar to gtk_widget_set_uposition(). that is: -1=unspecified,
-2=leave untouched. changed the width/height field of GtkWidgetAuxInfo
to be gint16s rather than guint16s, since that's what the code expected
(for a long time actually).
* gtk/gtkviewport.c:
added support for object arguments: GtkViewport::shadow_type,
GtkViewport::vadjustment and GtkViewport::hadjustment.
(gtk_viewport_add): chain gtk_bin_add.
* gtk/gtkscrolledwindow.c:
added support for object argument: GtkScrolledWindow::hscrollbar_policy,
GtkScrolledWindow::vscrollbar_policy and GtkScrolledWindow::viewport.
Tim Janik [Wed, 15 Jul 1998 12:25:43 +0000 (12:25 +0000)]
pass only the signal_id to this function. it needs to relookup the signal
Wed Jul 15 14:20:10 1998 Tim Janik <timj@gtk.org>
* gtk/gtksignal.c (gtk_signal_real_emit): pass only the signal_id
to this function. it needs to relookup the signal multiple times.
this fixes spurious segfaults that could show up with signal intensive
code, when classes are created during a pending emission (which is
normal case).
Tim Janik [Tue, 14 Jul 1998 07:40:15 +0000 (07:40 +0000)]
display the current cursor name.
Tue Jul 14 06:25:46 1998 Tim Janik <timj@gtk.org>
* gtk/testgtk.c (create_cursors): display the current cursor name.
* gtk/gtktypeutils.h:
* gtk/gtktypeutils.c (gtk_type_init_builtin_types): cleanups to
the fundamental type definition. the GtkArg union now only covers
the required storage types for argument values. adapted the GTK_VALUE_*
macro set. handle GTK_TYPE_UCHAR. definiton of type set bounds macros.
* gtk/gtkargcollector.c (gtk_arg_collect_value): clean ups of the
code, fixed int/float collection. fixed a bug in the sanity checking
code for GtkObjects.
* gtk/gtkbindings.c (gtk_binding_entry_add_signal): fixups of the
argument collection code.
* gtk/gtksignal.c (gtk_signal_collect_params): use the type set bounds
macros when distinguishing the return type.
George Lebl [Mon, 13 Jul 1998 19:43:33 +0000 (19:43 +0000)]
check if old_parent_private is not NULL before trying to g_list_remove the
Mon Jul 13 12:42:51 1998 George Lebl <jirka@5z.com>
* gdk/gdkwindow.h: (gdk_window_reparent) check if old_parent_private
is not NULL before trying to g_list_remove the child. A patch
from Michael Lausch.
Tim Janik [Sun, 12 Jul 1998 04:13:02 +0000 (04:13 +0000)]
call the base classes' object_init_func for derived objects with the
Sun Jul 12 05:59:26 1998 Tim Janik <timj@gtk.org>
* gtk/gtktypeutils.c (gtk_type_new): call the base classes'
object_init_func for derived objects with the object's ->klass field
still pointing to the corresponding base class, otherwise overridden
class functions could get called with partly-initialized objects.
(change from Tero Pulkkinen <terop@modeemi.cs.tut.fi>).
Tim Janik [Fri, 10 Jul 1998 11:36:26 +0000 (11:36 +0000)]
strictly check for the associated get/set arg functions. if an arg is
Fri Jul 10 13:29:53 1998 Tim Janik <timj@gtk.org>
* gtk/gtkcontainer.c:
(gtk_container_arg_get):
(gtk_container_arg_set):
* gtk/gtkobject.c:
(gtk_object_arg_set):
(gtk_object_arg_get):
strictly check for the associated get/set arg functions. if an
arg is readable or writable, the associated class functions *must*
be provided.
Tim Janik [Fri, 10 Jul 1998 03:59:44 +0000 (03:59 +0000)]
collect the arguments with gtk_container_child_args_collect, not through
Fri Jul 10 05:51:44 1998 Tim Janik <timj@gtk.org>
* gtk/gtkcontainer.c:
(gtk_container_add_with_args): collect the arguments with
gtk_container_child_args_collect, not through gtk_object_args_collect.
(gtk_container_child_set): likewise.
add in just enough support for 1 or 4 bits so that it seems to allow
Thu Jul 9 21:03:19 CDT 1998 Shawn T. Amundson <amundson@gtk.org>
* gtk/gtkpreview.c: add in just enough support for 1 or 4 bits
so that it seems to allow normal operation in those modes,
although you will only see black. 1 bit untested
Tim Janik [Thu, 9 Jul 1998 23:00:34 +0000 (23:00 +0000)]
deleted most of the argument handling code, since that is now implemented
Fri Jul 10 00:02:04 1998 Tim Janik <timj@gtk.org>
* gtk/gtkcontainer.h:
* gtk/gtkcontainer.c: deleted most of the argument handling code, since
that is now implemented in gtkarg.c. similar to gtk_object_args_collect,
we now export a new function gtk_container_child_args_collect().
for consistency with the object arguments, a few functions got renamed:
gtk_container_child_arg_get -> gtk_container_child_get,
gtk_container_child_arg_set -> gtk_container_child_set,
gtk_container_child_arg_getv -> gtk_container_child_getv,
gtk_container_child_arg_setv -> gtk_container_child_setv,
gtk_container_add_with_argv -> gtk_container_addv.
note, gtk_container_add_with_args() remained, because its equivalent
would be gtk_container_add(,, ...) which would break all existing code.
(gtk_container_add_child_arg_type): similar to gtk_object_add_arg_type,
we expect the `arg_name' argument to be a const static string now.
(gtk_container_get_child_arg_type): function removed.
Thu Jul 9 07:03:04 1998 Tim Janik <timj@gtk.org>
* gtk/gtkargcollector.c: new file which holds gtk_arg_collect_value().
this is a static inline function that collects command line arguments
from a va_list. this file can just be included in all places that
need this functionality.
* gtk/gtkarg.h:
* gtk/gtkarg.c: new files which implement most of the argument
handling stuff from gtkobject.c. also collected a few more
gtk_arg_* utility functions from else places.
* gtk/gtkobject.h:
* gtk/gtkobject.c: moved most of the argument handling code into
gtkarg.c. we now export gtk_object_args_collect() as a non-public
method with a blind va_list pointer.
(gtk_object_add_arg_type): the `arg_name' argument is required to be a
const static string now.
(gtk_object_get_arg_type): function got removed.
* gtk/gtkwidget.c:
(gtk_widget_set):
(gtk_widget_new): adaptions for gtk_object_args_collect().
* gtk/gtktypeutils.c (gtk_type_init_builtin_types): changed the internal
fundamental type name so as to have a valid prefix, e.g. "bool"->
"gboolean", "string"->"GtkString" and somesuch, left "void" as is,
though that should probably be something like GtkNone since the
type itself is called GTK_TYPE_NONE.
even the internal type names need to avoid name clashes and must live
in their own namespace, several code portions rely on that.
we should relly have typedefs such as typedef gchar* GtkString; so the
fundamental type names can be used for code dumpers just like with all
the Gtk/Gdk types.
Tim Janik [Wed, 8 Jul 1998 14:12:27 +0000 (14:12 +0000)]
handle GTK_RELIEF_HALF, which was introduced to keep either
Wed Jul 8 15:31:28 1998 Tim Janik <timj@gtk.org>
* gtk/gtkbutton.c: handle GTK_RELIEF_HALF, which was introduced to
keep either GTK_STATE_NORMAL for relief buttons (GTK_RELIEF_NONE) or
to still honour GTK_WIDGET_STATE(), e.g. for coloured buttons
(GTK_RELIEF_HALF).
(gtk_button_paint): set the background for prelighted buttons with
GTK_STATE_NORMAL only if GTK_RELIEF_NONE.
(gtk_button_draw_focus): if GTK_RELIEF_NONE, use GTK_STATE_NORMAL for
the shadow type (otherwise we get strange shadows for coloured buttons).
(gtk_button_set_relief): queue a redraw after changing
the relief.
(gtk_button_class_init):
(gtk_button_set_arg):
(gtk_button_get_arg): added object argument "GtkButton::relief".
Tim Janik [Tue, 7 Jul 1998 01:25:27 +0000 (01:25 +0000)]
totaly binary incomnpatible change:
Mon Jul 6 18:30:48 1998 Tim Janik <timj@gtk.org>
* gtk/gtkbutton.h:
* gtk/gtkbutton.c: GtkButtons are now derived from GtkBin. we mirror
GTK_BIN (button)->child to button->child to keep source compatibility.
* gtk/gtkoptionmenu.c:
* gtk/gtkclist.c:
* gtkcheckbutton.c: don't refer to button->child, but
GTK_BIN (button)->child instead.
Tim Janik [Tue, 30 Jun 1998 12:34:36 +0000 (12:34 +0000)]
fixed lookup order for rc styles. the latest defined rc set pattern is
Tue Jun 30 14:30:37 1998 Tim Janik <timj@gtk.org>
* gtk/gtkrc.c: fixed lookup order for rc styles. the latest defined
rc set pattern is favoured, this is needed for overriding of style
patterns with subsequently parsed rc files.
Tim Janik [Tue, 30 Jun 1998 08:20:45 +0000 (08:20 +0000)]
ok, this enables the possibility for key-release bindings.
but nearly no widget really cares to set GDK_KEY_RELEASE_MASK,
which needs to be enabled in order for release bindings to work.
Tue Jun 30 09:24:40 1998 Tim Janik <timj@gtk.org>
* gtk/gtkaccelgroup.c (gtk_accelerator_parse): parse "<Release>"
modifier as well.
(gtk_accelerator_name): add "<Release>" to the accelerator name if
neccessary.
* gtk/gtkbindings.c (BINDING_MOD_MASK): take BINDING_MOD_MASK into
account.
* gtk/gtkwidget.c (gtk_widget_real_key_release_event): implemented
default handler for key-release events, which checks for activation
of GDK_RELEASE_MASK key bindings.
* gdk/gdktypes.h (enum): added GDK_RELEASE_MASK to GdkModifierType.
Tim Janik [Tue, 30 Jun 1998 08:02:53 +0000 (08:02 +0000)]
implemented default handler for key-release events, which checks for
Tue Jun 30 09:24:40 1998 Tim Janik <timj@gtk.org>
* gtk/gtkwidget.c (gtk_widget_real_key_release_event): implemented
default handler for key-release events, which checks for activation
of GDK_RELEASE_MASK key bindings.
* gdk/gdktypes.h (enum): added GDK_RELEASE_MASK to GdkModifierType.
Tim Janik [Sun, 28 Jun 1998 07:46:10 +0000 (07:46 +0000)]
call the base class init fucntions from all parent types upon class
Sun Jun 28 04:29:10 1998 Tim Janik <timj@gtk.org>
* gtk/gtktypeutils.c (gtk_type_class_init): call the base class init
fucntions from all parent types upon class initialization.
* gtk/gtkcontainer.c:
(gtk_container_get_type): announce gtk_container_base_class_init to
the type system.
(gtk_container_base_class_init): new function to feature base class
initialization.
(gtk_container_get_child_arg):
(gtk_container_set_child_arg): call the GtkContainerClass get_child_arg
and set_child_arg methods of the class indicated through the argument
name.
* gtk/gtkobject.c:
(gtk_object_base_class_init): new function to feature base class
initialization.
(gtk_object_init_type): announce gtk_object_base_class_init to the type
system.
(gtk_object_class_init): setup the get_arg and set_arg pointers for
GtkObjectClass.
(gtk_object_setv):
(gtk_object_getv): call the GtkObjectClass get_arg and set_arg methods,
instead of bothering the type system with this.
* gtk/gtkaccellabel.c:
* gtk/gtkbutton.c:
* gtk/gtkradiobutton.c:
* gtk/gtktable.c:
* gtk/gtktogglebutton.c:
* gtk/gtktipsquery.c:
* gtk/gtkbox.c:
* gtk/gtkpacker.c:
* gtk/gtkwidget.c:
* gtk/gtkwindow.c:
* gtk/gtkframe.c:
* gtk/gtkmisc.c:
* gtk/gtklabel.c: set the object_class->{g|s}et_arg pointers to the
corresponding gtk_*_{g|s]et_arg functions and updated the gtk_*_get_type
functions wrt GtkTypeInfo initialization. changed a lot of the set/get
arg functions to take a GtkObject argument.
add section on GtkCList widget, contributed by Stefan Mars
Thu Jun 25 07:53:51 BST 1998 Tony Gale <gale@gtk.org>
* docs/gtk_tut.sgml: add section on GtkCList widget, contributed by Stefan Mars <mars@lysator.liu.se>
* examples/clist/clist.c examples/clist/Makefile: example code
for GtkCList widget from the Tutorial
Tim Janik [Wed, 24 Jun 1998 17:15:05 +0000 (17:15 +0000)]
do not avoid to queue for a resize if the container is not visible, we
Wed Jun 24 16:38:02 1998 Tim Janik <timj@gtk.org>
* gtk/gtkbin.c (gtk_bin_remove): do not avoid to queue for a resize
if the container is not visible, we might be a toplevel! this holds
for all other base container implementations as well, that are
candidates to derive toplevels from. in general the resizing code will
care about visibility itself.
* gtk/gtkmenushell.c (gtk_menu_shell_remove): likewise.
* gtk/gtkbox.c (gtk_box_remove): likewise.
* gtk/gtkwindow.c (gtk_window_move_resize): save ->use_uposition around
gtk_window_set_hints(), since we haven't calculated the new position
yet.
* gdk/gdkwindow.c (gdk_window_get_pointer): take care that *x, *y and
*mask are alway initialized in a sane way, regardless of the return
value.