Tue Feb 8 09:38:29 2000 Tim Janik <timj@gtk.org>
* gtk/gtkwidget.c:
(gtk_widget_unrealize): guard widget access with ref/unref
around signal emission.
(gtk_widget_hide): same here, but also check its destroyed
state before queueing a resize.
Tue Feb 8 03:05:55 2000 Tim Janik <timj@gtk.org>
* gtk/gtkstyle.c (gtk_style_new): use gtk_default_prelight_bg instead
of gtk_default_insensitive_bg as insensitive base color.
* gtk/gtktext.c (gtk_text_style_set): set the background color according
to the widget's state.
(gtk_text_realize): same here.
(gtk_text_state_changed): same here.
(draw_bg_rect): compare background color against base[] from
GTK_WIDGET_STATE (text) not GTK_STATE_NORMAL.
Mon Feb 7 04:01:55 2000 Tim Janik <timj@gtk.org>
* gtk/gtkmain.c (gtk_main_do_event): ignore delete events when
grabs are active, unless a delete event is send to the toplevel
of the currently grab holding widget.
Fri Feb 4 19:40:50 2000 Tim Janik <timj@gtk.org>
* configure.in: up version number to 1.2.7, interface age 2, binary
age 7. depend on glib 1.2.7.
* NEWS: updates.
* gtk/gtkscrolledwindow.c (gtk_scrolled_window_size_request): Don't
ever make the scrolled-window requisition ever depend on the
visibility state of the scrollbars for the AUTOMATIC policy.
This breaks the GTK+ requisition model, and causes loops.
Fri Feb 4 15:09:12 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkstyle.c (gtk_style_new): Set the charset explicitely for
the default font to avoid problems with XFree86-4.0 where the
default charset is iso10646-1, not iso8859-1.
Thu Feb 3 14:46:01 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkmenushell.c (gtk_menu_shell_button_release): Set
the activate_time to 0, so that we handle a quick
press/release press/release properly and don't suppress
the second release. (Red Hat bug #7545)
Sun Jan 30 20:10:52 2000 Lars Hamann <lars@gtk.org>
A few more fixes for bug #5487, #2051, #2677.
* gtk/gtkclist.c :
(gtk_clist_button_press): Reset clist->anchor to -1 if event->type is
not GDK_BUTTON_PRESS.
(resync_selection): resync only if selection_mode is
GTK_SELECTION_EXTENDED
* gtk/gtkctree.c (resync_selection): same here
Sun Jan 30 12:29:20 2000 Owen Taylor <otaylor@redhat.com>
* gdk/gdk.h gtk/gtkmenu.h: Removed duplicate prototypes for
gdk_drag_get_selection() and gtk_menu_detach().
[ From Jeroen Ruigrok/Asmodai ]
2000-01-25 Havoc Pennington <hp@pobox.com>
* gdk/gdkwindow.c (gdk_window_set_back_pixmap): parent_relative
arg is boolean
* gdk/gdkselection.c (gdk_selection_owner_set): return boolean
* gdk/gdkrectangle.c (gdk_rectangle_intersect): return boolean
* gdk/gdkproperty.c (gdk_property_get): return boolean
* gdk/gdkinput.c (gdk_input_set_mode): return boolean
* gdk/gdkim.c (gdk_im_ready): return boolean
* gdk/gdkgc.c (gdk_gc_set_exposures): gboolean arg
* gdk/gdkfont.c (gdk_font_equal): return gboolean (mild
glib clash, should fix glib)
* gdk/gdkevents.c (gdk_set_show_events): take gboolean arg
(gdk_get_show_events): return gboolean, and canonicalize
to TRUE/FALSE
* gdk/gdkcolor.c (gdk_colormap_new): gboolean flag whether the
colormap is private
(gdk_colors_alloc): gboolean whether to be contiguous
(gdk_color_equal): return gboolean since we are a predicate
and not a qsort() (this looks semi-wrong due to glib breakage,
IMO glib should be fixed)
* gdk/gdk.c (gdk_set_use_xshm): take gboolean arg
(gdk_get_use_xshm): return gboolean
(gdk_pointer_is_grabbed): return gboolean
* gdk/gdk.h: Change prototypes to match all the above changes,
and re-run egtk-format-protos as required.
Fri Jan 28 12:28:17 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkctree.c (resync_selection):
* gtk/gtkclist.c (resync_selection):
Return immediately if clist->drag_pos < 0. This is a workaround
for the corrupt state that the clist gets into when a
GDK_2BUTTON_PRESS occurs. See note in gtk_clist_button_press().
Bug #5487, #2051, #2677; fix suggested by David Helder and T. Alexander Popiel.
* gtk/gtknotebook.c (gtk_notebook_size_request): Fix uninitialized
variable from recent commit.
Thu Jan 27 18:00:55 2000 Tim Janik <timj@gtk.org>
* gtk/Makefile.am: prefix all autogenerated source that get build in
$(srcdir) with $(srcdir)/, so make doesn't assume they got generted
in builddir. since we subsequently cd into srcdir for autogeneration,
the paths have to be stripped from the target file names, thusly we
use $(@F) as target names now.
put a comment about configure.in's --disable-rebuilds option,
which can be used for non-writable source directories, for development
setups though, srcdir has to be *writable*.
Thu Jan 27 00:15:03 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkdnd.c (gtk_drag_begin gtk_drag_get_event_actions):
Finish the job of allowing event to be NULL.
(Fixes bug #4283, reported by Chris Blizzard)
* gtk/gtkentry.c (entry_adjust_scroll): When calculating
things so that the cursor appears on screen, properly
take into account INNER_BORDER.
(Fixes bug #4754, reported by Antonio Campos)
Wed Jan 26 23:12:28 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtknotebook.c (gtk_notebook_map): Don't
show the tab_label unless it itself is visible.
gtk/gtknotebook.c (gtk_notebook_size_request):
Do a better job of making sure that the visibility
of the tab label corresponds to whether it should
be mapped or not.
Wed Jan 26 21:17:03 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkoptionmenu.c (gtk_option_menu_position):
Use the requisition, not the allocation, since the
allocation has not necessarily been computed yet.
(Pointed out by Eugene Osintsev)
Wed Jan 26 19:44:25 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkstyle.c (gtk_style_new): Dont' set
style/base[GTK_STATE_INSENSITIVE] both to
gtk_default_insensitive_bg!
(Bug #2187, reported by Jonathan Blandford)
* gtk/gtkaccelgroup.c (gtk_accelerator_valid):
Add Alt_L, Alt_R to list of invalid accelerators.
(Bug #3736, reported by Vlad Harchev)
Wed Jan 26 19:01:56 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkfilesel.c (open_ref_dir): Fix several
bugs which occured after an attempt to open
invalid home directory left cmpl_state->reference_dir == NULL.
- completion on files in home directory didn't work
- completion on an empty string caused segfault
(Bug #3768, reported by Steve Ratcliffe)
* gtk/gtkscale.c (gtk_scale_get_value_width): Fix
cut and paste error that was causing scales to
be incorrectly positioned.
(Bug #2956,
patch from Steve Ratcliffe <steve@parabola.demon.co.uk>)
Wed Jan 26 18:06:07 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwidget.c (gtk_widget_queue_clear_area): Ignore
queue_clear on NO_WINDOW widgets during a reparent,
since at that point the window and widget heirarchies
are out of sync. This stops crashing in some cases
(the scrolled window reparent test, for instance), and
_probably_ won't cause drawing errors.
* gtk/testgtk.c: Make the reparenting-a-scrolled-window
test do what it was supposed to do and be robust against
window closings, etc. (Bug #2443)
Wed Jan 26 16:56:54 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkentry.c: Fix return values on mouse events.
(Bug #2686, Sky <seb_sky@yahoo.com>)
* gtk/genmarshal.pl: Fix up handling of FOREIGN.
(pointed out by George Lebl)
* gdk/gdkcolor.c (gdk_color_copy): Made const-safe.
(From Ettore Perazzoli <ettore@helixcode.com>)
Mon Jan 24 10:44:48 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkcalendar.c (gtk_calendar_main_button): Handle
clicks between rows or outside calender area correctly.
(Patch from Damon Chaplin)
* gtk/gtkclist.c (gtk_clist_set_row_data_full):
* gtk/gtkctree.c (gtk_ctree_node_set_row_data_full):
Call destroy function when overwriting existing data.
(Pointed out by Damon Chaplin)
Tue Jan 25 09:55:41 2000 Owen Taylor <otaylor@redhat.com>
* gdk/gdk.c: Remove useless #include of <X11/Xmu/WinUtil.h>,
which was causing problems on various systems with Xmu.
This hasn't been needed for a very long time.
(Fixes#1185 3167)
Sun Jan 23 22:27:36 2000 Owen Taylor <otaylor@redhat.com>
* gdk/gdki18n.h: Include <ctype.h> when defining
gdk_isw* in terms of is* as a fallback.
(Bug #4106 - Dan Winship <danw@MIT.EDU>)
Sun Jan 23 22:12:36 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkviewport.c (gtk_viewport_size_allocate): Don't
realize the widget when it is size allocated!
(old, old bug)
* gtk/gtkviewport.c (gtk_viewport_adjustment_value_changed):
Removed unused call to gdk_window_get_size() that
showed up when the above was fixed.
Sun Jan 23 20:52:40 2000 Owen Taylor <otaylor@redhat.com>
* INSTALL: Added short note about setting LDFLAGS
and CPPFLAGS to find libraries not in the default
install path. Fixed env/setenv confusion for csh.
Sat Jan 22 15:44:30 2000 Owen Taylor <otaylor@redhat.com>
* gtk/gtkwidget.c (gtk_widget_reparent): Correctly
fix up widget->window when the widget is a
NO_WINDOW container widget.
Sat Jan 22 12:40:48 2000 Owen Taylor <otaylor@redhat.com>
* gdk/gdkpixmap.c (gdk_pixmap_seek_string): Rewrite
to account for the fact that feof() does _not_
return TRUE on errors, and thus avoid infinite loops
when trying to use gdk_pixmap_create_from_xpm()
on unreadable values.
Fri Jan 21 16:24:08 2000 Owen Taylor <otaylor@redhat.com>
* gdk/gdkgc.c (gdk_gc_set_dashes): Change from
gdk_gc_set_dashes to take gint8 instead of gchar to
make it clearer that it is _not_ a NULL terminated string.
* gdk/gdk.h gdk/gdkfont.c gdk/gdkgc.c gdk/gdkselection.c
gdk/gdkwindow.c gtk/gtkprogress.[ch] gtk/gtkthemes.[ch]
gtk/gtktreeitem.[ch] gtk/gtkwidget.[ch]: Constify string
arguments.
Thu Dec 30 04:51:05 1999 Tim Janik <timj@gtk.org>
* gtk/gtkvscale.c (gtk_vscale_pos_background):
* gtk/gtkhscale.c (gtk_hscale_pos_background): when calculating the
background size relative to our allocation, guard against small
allocations, we may have not yet been size allocated.
Fri Dec 24 03:41:33 1999 Tim Janik <timj@gtk.org>
* gtk/Makefile.am: reverted sopwith's changes to Makefile.am,
the builddir != srcdir case was already handled fine, and his
"fix" would produce multiple copies of the built sources.
for development you just need a writable srcdir.
