]> Pileus Git - ~andy/gtk/log
~andy/gtk
26 years agoFrom Patrice Fortier <Patrice.Fortier@aquarel.fr>
Owen Taylor [Tue, 17 Feb 1998 05:07:51 +0000 (05:07 +0000)]
From Patrice Fortier <Patrice.Fortier@aquarel.fr>

Tue Feb 17 00:06:26 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkpaned.c, gtk/gtk[hv]paned.c: (gtk-fortier-980215-0)
  From Patrice Fortier <Patrice.Fortier@aquarel.fr>

  Add a widget->window for paned windows, so they
  can be reparented properly.

26 years agoFrom Patrice Fortier <Patrice.Fortier@aquarel.fr>
Owen Taylor [Tue, 17 Feb 1998 04:48:50 +0000 (04:48 +0000)]
From Patrice Fortier <Patrice.Fortier@aquarel.fr>

Mon Feb 16 23:47:09 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkviewport.c: (gtk-fortier-980130-0)
  From Patrice Fortier <Patrice.Fortier@aquarel.fr>

  Eliminate some extra expose events for viewports.
  (To be conservative, I set the user's event mask for
   both widget->window, and viewport->bin_window  -owt)

26 years agoMake timeout and idle insertion code use g_list_insert_sorted().
Owen Taylor [Tue, 17 Feb 1998 04:39:47 +0000 (04:39 +0000)]
Make timeout and idle insertion code use g_list_insert_sorted().

Mon Feb 16 23:12:14 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkmain.c: Make timeout and idle insertion code
  use g_list_insert_sorted().

26 years agoChanged function so elements are always inserted, even if they compare
Owen Taylor [Tue, 17 Feb 1998 04:03:56 +0000 (04:03 +0000)]
Changed function so elements are always inserted, even if they compare

Mon Feb 16 23:05:06 1998  Owen Taylor  <owt1@cornell.edu>

* glist.c (g_list_insert_sorted): Changed function
  so elements are always inserted, even if they compare
  equal with another.

26 years ago\n
Tim Janik [Tue, 17 Feb 1998 02:19:30 +0000 (02:19 +0000)]
\n

26 years agoUse gtk_signal_default_marshaller, for expand/collapse signals.
Owen Taylor [Mon, 16 Feb 1998 22:34:30 +0000 (22:34 +0000)]
Use gtk_signal_default_marshaller, for expand/collapse signals.

Mon Feb 16 17:35:17 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtktreeitem.c: Use gtk_signal_default_marshaller,
  for expand/collapse signals.

26 years agoD&D fixes
Elliot Lee [Mon, 16 Feb 1998 22:04:52 +0000 (22:04 +0000)]
D&D fixes

26 years agogtkmain forgot a return value.
Tim Janik [Sat, 14 Feb 1998 05:37:53 +0000 (05:37 +0000)]
gtkmain forgot a return value.
few fixes with resizing behaviour and non opaque moves
where configure events would exceed te number of calls
to gdk_window_resize.
-timj

26 years agoFixed some incorrect information.
Owen Taylor [Fri, 13 Feb 1998 23:00:54 +0000 (23:00 +0000)]
Fixed some incorrect information.

26 years agocheck for clist !REALIZED. (draw_rows): bail out if the widget isn't
Tim Janik [Fri, 13 Feb 1998 18:41:18 +0000 (18:41 +0000)]
check for clist !REALIZED. (draw_rows): bail out if the widget isn't

Fri Feb 13 19:08:51 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkclist.c (check_exposures): check for clist !REALIZED.
        (draw_rows): bail out if the widget isn't drawable.
        (hadjustment_value_changed):
        (vadjustment_value_changed): bail out if clist isn't drawable.

        * gdk/gdk.c (graphics_expose_predicate): check for private != NULL.
        (gdk_event_get_graphics_expose): check for window != NULL.

26 years agoRealize window before creating xpm so it gets the right colormap/visual,
Owen Taylor [Fri, 13 Feb 1998 05:33:17 +0000 (05:33 +0000)]
Realize window before creating xpm so it gets the right colormap/visual,

Fri Feb 13 00:33:26 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/testgtk.c (shape_create_icon): Realize window before
  creating xpm so it gets the right colormap/visual,
  instead of the default one.

26 years agoChange style->white to style->base[] where appropriate. Change the default
Owen Taylor [Fri, 13 Feb 1998 05:26:33 +0000 (05:26 +0000)]
Change style->white to style->base[] where appropriate. Change the default

* gtk/gtkentry.c gtk/gtkinputdialog.c gtk/gtklist.c
  gtk/gtklistitem.c gtktree.c gtk/gtktreeitem.c testgtkrc:

  Change style->white to style->base[] where appropriate.
  Change the default style to make this apparent. (Yes, it's
  ugly... it can be removed later)

26 years agogtk/gtksignal.[ch] gtk/gtkmain.[ch] gtk/gtkcontainer.[ch]
Owen Taylor [Fri, 13 Feb 1998 05:19:06 +0000 (05:19 +0000)]
gtk/gtksignal.[ch] gtk/gtkmain.[ch] gtk/gtkcontainer.[ch]

Thu Feb 12 23:59:49 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtksignal.[ch] gtk/gtkmain.[ch] gtk/gtkcontainer.[ch]
  gtk/gtkselection.[ch] gdk/gdk.[ch] gtktypeutils.h

    Replaced all _interp functions with _full functions.
  (_interp functions left in for compatibility until 0.99.5).
  Difference: _full functions take _both_ a C-language callback and a
  marshaller, and simply ignore the C-language callback
  if the marshaller is present. This allows the destroy notification
  to be used without marshalling.

  gtk_selection_add_handler[_full]() regularized to agree
  with other callbacks.

  Also, added gtk_input_add_full() to the header file.
  (gtk_input_add_interp() was never there)

* gtk/gtkmain.[ch] gtk/gtkwidget.c: Idle functions are now prioritized.
  Added new function gtk_idle_add_priority to create
  an idle with a specified priority (default is zero)
  constants #defined - GTK_PRIORITY_[HIGH/INTERNAL/DEFAULT/LOW]
  (redraws, resizes run at GTK_PRIORITY_INTERNAL)

* gtk/gtkentry.c gtk/testselection.c: changes to keep up with change
     to gtkselection.c.

26 years agoAdd a DestroyNotify field, and gtk_object_set_data_full() to match.
Owen Taylor [Fri, 13 Feb 1998 05:11:16 +0000 (05:11 +0000)]
Add a DestroyNotify field, and gtk_object_set_data_full() to match.

Tue Feb 10 15:01:44 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkobject.c gtk/gtkobject.h: Add a DestroyNotify
field, and gtk_object_set_data_full() to match.

* gtk/gtkobject.c (gtk_object_finalize): ObjectData structures
were being added to a free list, then forgotten about. Just
rely on GMemChunk instead.

26 years agochange logged.
Owen Taylor [Fri, 13 Feb 1998 03:47:43 +0000 (03:47 +0000)]
change logged.

26 years agoremoved deprecated g_string_equal and g_string_hash.
Owen Taylor [Fri, 13 Feb 1998 03:47:25 +0000 (03:47 +0000)]
removed deprecated g_string_equal and g_string_hash.

Thu Feb 12 22:48:11 1998  Owen Taylor  <owt1@cornell.edu>

* gstring.c glib.h: removed deprecated g_string_equal
  and g_string_hash.

26 years agoFixed unrealize handler to comply with the documentation (which mentions
Federico Mena Quintero [Thu, 12 Feb 1998 03:02:14 +0000 (03:02 +0000)]
Fixed unrealize handler to comply with the documentation (which mentions

2018-02-11  Federico Mena Quintero  <federico@nuclecu.unam.mx>

* gtk/gtkclist.c:
* gtk/gtkentry.c:
* gtk/gtkfixed.c:
* gtk/gtkhandlebox.c:
* gtk/gtknotebook.c:
* gtk/gtkpaned.c:
* gtk/gtkpreview.c:
* gtk/gtkrange.c:
* gtk/gtkruler.c:
* gtk/gtktext.c:
* gtk/gtkviewport.c:
Fixed unrealize handler to comply with the
documentation (which mentions the right thing to do).

26 years agoSpelling fixups and minor changes.
Owen Taylor [Thu, 12 Feb 1998 02:40:30 +0000 (02:40 +0000)]
Spelling fixups and minor changes.

26 years agosmall fixup, so that the clist doesn't draw if it is
Tim Janik [Wed, 11 Feb 1998 05:07:36 +0000 (05:07 +0000)]
small fixup, so that the clist doesn't draw if it is
not realized.
-timj

26 years agoReally minor cosmetic fix that just niggled me. :P
Manish Singh [Wed, 11 Feb 1998 01:08:42 +0000 (01:08 +0000)]
Really minor cosmetic fix that just niggled me. :P

-Yosh

26 years agobacked out the section "Gnits to care about". new file, kinda developers
Tim Janik [Wed, 11 Feb 1998 00:40:20 +0000 (00:40 +0000)]
backed out the section "Gnits to care about". new file, kinda developers

Wed Feb 11 00:18:31 1998  Tim Janik  <timj@gimp.org>

        * docs/refcounting.txt: backed out the section "Gnits to care about".
        * docs/developers.txt: new file, kinda developers FAQ.

26 years agonew function gtk_signal_handler_pending() returning the id of the next
Tim Janik [Tue, 10 Feb 1998 23:49:15 +0000 (23:49 +0000)]
new function gtk_signal_handler_pending() returning the id of the next

Wed Feb 11 00:18:31 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtksignal.h:
        * gtk/gtksignal.c: new function gtk_signal_handler_pending() returning
          the id of the next handler pending for that signal or 0.
          put struct GtkHandler back into gtksignal.c along with
          gtk_signal_get_handlers.

26 years agoAdd check to see if the C library's iswalnum can actually be used. (Not
Owen Taylor [Tue, 10 Feb 1998 18:18:19 +0000 (18:18 +0000)]
Add check to see if the C library's iswalnum can actually be used. (Not

Tue Feb 10 13:04:36 1998  Owen Taylor  <owt1@cornell.edu>

* configure.in: Add check to see if the C library's
  iswalnum can actually be used. (Not true for
  Linux libc-5.4.38)

26 years agook, there have been several severe bugs in the signal handler referencing
Tim Janik [Tue, 10 Feb 1998 06:53:08 +0000 (06:53 +0000)]
ok, there have been several severe bugs in the signal handler referencing

Tue Feb 10 07:12:07 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtksignal.h:
        * gtk/gtksignal.c:
        ok, there have been several severe bugs in the signal handler
        referencing and ->next connection stuff. these bugs caused
        invokations of handlers that are disconnected and - worse -
        destroyed already. invokation of *destroyd* handlers mean:
        anything can be executed , because the handler structure can just
        as well be realocated.
        at the cost of an extra ->prev field per handler we should have a
        reasonable stable system now, because of the various places that
        can cause a handler to be disconnected (*any* handler invokation can
        cause *any* or *all* handlers to be disconnected, there is no way
        around a doubly linked list, actually handler disconnection has never
        worked correctly because of this.
        handlers are connected together via a *doubly* linked list now, and it
        is *not* valid to remove a handler out of this list untill all its
        references have been droped, i.e. handler->ref_count==0.
        to prevent emissions of disconnected but still referenced handlers,
        disconnected handlers are simply marked as blocked and get an id of 0
        which is an invalid signal handler id.
        the handler->id has been changed to have 28 significant bits (using
        alignment gaps), since 65536 (old range: guint16) signal connections
        (as a total) can easily be reached by complex applications.
        this whole handler thingy is at least as tedious as writing doubly
        linked list implementations ;)

26 years agoCheck for the RESIZE_NEEDED flag and remove the widget from the list.
Owen Taylor [Tue, 10 Feb 1998 04:17:10 +0000 (04:17 +0000)]
Check for the RESIZE_NEEDED flag and remove the widget from the list.

Mon Feb  9 23:08:16 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkwidget.c (gtk_widget_unparent): Check for
  the RESIZE_NEEDED flag and remove the widget from the list.
  Remove the check from gtk_widget_destroy (no longer needed).

* Unrealize widget _before_ calling "destroy" signal, and
  unset VISIBLE flag.

* Unrealize child widgets _after_ unrealizing parent to improve
  visual appearance.

26 years agono longer a private routine; gtksignal.h: export gtk_signal_get_handlers
Miguel de Icaza [Mon, 9 Feb 1998 22:44:34 +0000 (22:44 +0000)]
no longer a private routine; gtksignal.h: export gtk_signal_get_handlers

Mon Feb  9 16:42:21 1998  Miguel de Icaza  <miguel@nuclecu.unam.mx>

* gtk/gtksignal.c (gtk_signal_get_handlers): no longer a private
routine;
gtksignal.h: export gtk_signal_get_handlers and GtkHandlers type.

26 years agonew fundtion to return the widget that the menu is attached to.
Tim Janik [Sun, 8 Feb 1998 19:00:01 +0000 (19:00 +0000)]
new fundtion to return the widget that the menu is attached to.

Sun Feb  8 07:06:54 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkmenu.c (gtk_menu_get_attach_widget): new fundtion to return
          the widget that the menu is attached to.

26 years agoNew API functions: gtk_clist_get_cell_type gtk_clist_get_text
Jay Painter [Sun, 8 Feb 1998 07:55:11 +0000 (07:55 +0000)]
New API functions: gtk_clist_get_cell_type gtk_clist_get_text

New API functions:
  gtk_clist_get_cell_type
  gtk_clist_get_text
  gtk_clist_get_pixmap
  gtk_clist_get_pixtext
  gtk_clist_find_row_from_data

26 years agouse a mask of 0xffffffff instead of 0xffff, since there are 22 flags
Owen Taylor [Sat, 7 Feb 1998 17:07:14 +0000 (17:07 +0000)]
use a mask of 0xffffffff instead of 0xffff, since there are 22 flags

Sat Feb  7 11:33:08 1998  Owen Taylor  <owt1@cornell.edu>

* gdk/gdkgc.c (gdk_gc_copy): use a mask of 0xffffffff
  instead of 0xffff, since there are 22 flags currently
  defined. (From: Jon Trowbridge <trow@emccta.com>)

  Changed to something better. ~((~1) << GCLastBit)
   (From: from Daniel Stephens <daniel@cheeseplant.org>)

26 years agoadded some additional consts in appropriate places to remove a warning
Owen Taylor [Sat, 7 Feb 1998 16:48:03 +0000 (16:48 +0000)]
added some additional consts in appropriate places to remove a warning

Sat Feb  7 11:48:09 1998  Owen Taylor  <owt1@cornell.edu>

* gstring.c gutils.c: added some additional consts in
  appropriate places to remove a warning

26 years agouse a mask of 0xffffffff instead of 0xffff, since there are 22 flags
Owen Taylor [Sat, 7 Feb 1998 16:34:40 +0000 (16:34 +0000)]
use a mask of 0xffffffff instead of 0xffff, since there are 22 flags

Sat Feb  7 11:33:08 1998  Owen Taylor  <owt1@cornell.edu>

* gdk/gdkgc.c (gdk_gc_copy): use a mask of 0xffffffff
  instead of 0xffff, since there are 22 flags currently
  defined. (From: Jon Trowbridge <trow@emccta.com>)

26 years agoinclude <ctype.h> for tolower()
Owen Taylor [Sat, 7 Feb 1998 16:14:13 +0000 (16:14 +0000)]
include <ctype.h> for tolower()

Sat Feb  7 11:15:54 1998  Owen Taylor  <owt1@cornell.edu>

* gutils.c: include <ctype.h> for tolower()

26 years agoHere's a small patch that corrects a prototype in gtkvpaned.h.
Tom Tromey [Sat, 7 Feb 1998 16:10:01 +0000 (16:10 +0000)]
Here's a small patch that corrects a prototype in gtkvpaned.h.

From: Tom Tromey <tromey@creche.cygnus.com>
Here's a small patch that corrects a prototype in gtkvpaned.h.

                                        -owt

26 years agoclean ups. -timj
Tim Janik [Sat, 7 Feb 1998 04:59:23 +0000 (04:59 +0000)]
clean ups. -timj

clean ups.
-timj

26 years agodon't wipe out window_private's x and y coordinates if the configure
Tim Janik [Sat, 7 Feb 1998 02:26:09 +0000 (02:26 +0000)]
don't wipe out window_private's x and y coordinates if the configure

Sat Feb  7 02:29:01 1998  Tim Janik  <timj@gimp.org>

        * gdk/gdk.c (gdk_event_translate): don't wipe out window_private's x
          and y coordinates if the configure notify is only about resizing,
          query the correct origin instead.

26 years agoA little cosmetic cleanup
Jay Painter [Thu, 5 Feb 1998 08:40:06 +0000 (08:40 +0000)]
A little cosmetic cleanup

26 years agoMenufactory now refs&sinks its children.
scott [Thu, 5 Feb 1998 04:10:43 +0000 (04:10 +0000)]
Menufactory now refs&sinks its children.
--sg

26 years agonew function to perform the same actions as gtk_list_remove_items, but
Tim Janik [Thu, 5 Feb 1998 03:53:41 +0000 (03:53 +0000)]
new function to perform the same actions as gtk_list_remove_items, but

Thu Feb  5 02:13:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtklist.h:
        * gtk/gtklist.c (gtk_list_remove_items_no_unref): new function
          to perform the same actions as gtk_list_remove_items, but
          supply the removed widgets with an additional reference count.

        * gtk/gtkmain.c (gtk_main_iteration_do): ignore events
          with event_widget == NULL, since they are bogus events
          from destroyed GdkWindows, exept for the case where
          event->type==GDK_PROPERTY_NOTIFY. Always handle expired
          timeout functions when returning from this function.

        * gtk/gtkwidget.c (gtk_widget_event): ignore GDK_EXPOSE events
          if event->window == NULL. Also, if this function couldn't handle
          the event for any reason (including failing assumptions), make
          the return value to look as if the event had been handled to
          avoid further processing (and warnings).

        * gtk/gtkwidget.h:
        * gtk/gtkwidget.c: remove gtk_widget_sink, because there is
          no point in providing such a function.

        * gdk/gdk.c (gdk_init): changed options `-name' and `-class'
          to `--name' and `--class', because the old names would
          confuse getopt(). these arguments have been introduced in the
          changes from gtk+970916 to gtk+970925 without a ChangeLog entry,
          changing argument names is painful, it would be nice if people
          would care about compatibility and consistency in the first place!

26 years agosection: Gnits to care about -------------------
Tim Janik [Thu, 5 Feb 1998 00:13:42 +0000 (00:13 +0000)]
section: Gnits to care about -------------------

section:
Gnits to care about
-------------------

-timj

26 years agodocumentation. -timj
Tim Janik [Wed, 4 Feb 1998 03:18:53 +0000 (03:18 +0000)]
documentation. -timj

documentation.
-timj

26 years agoindentations. -timj
Tim Janik [Wed, 4 Feb 1998 00:06:29 +0000 (00:06 +0000)]
indentations. -timj

indentations.
-timj

26 years agoRenamed widget_states.txt to widget_system.txt, and
Owen Taylor [Tue, 3 Feb 1998 23:31:21 +0000 (23:31 +0000)]
Renamed widget_states.txt to widget_system.txt, and
made some further revisions.

26 years agofor the emission of AFTER signals, fetch the objects signals via
Tim Janik [Tue, 3 Feb 1998 21:36:06 +0000 (21:36 +0000)]
for the emission of AFTER signals, fetch the objects signals via

Tue Feb  3 15:09:55 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtksignal.c (gtk_signal_real_emit): for the emission
          of AFTER signals, fetch the objects signals via
          gtk_signal_get_handlers again. some handlers might have
          been removed or added. not doing this would mess up the
          memchunk allocation of signal handlers (this had been
          triggered by multiple *_while_alive connections), bad, bad, bad!
          (gtk_handlers_run): do the referencing on signal handlers
          unconditionally, the invokation of AFTER handlers will now take
          care of modified lists.

        * gtk/gtksignal.h: added gtk_signal_connect_while_alive.

26 years agoOnly query the colormap for GRAYSCALE and PSEUDOCOLOR visuals, and don't
Owen Taylor [Tue, 3 Feb 1998 20:57:40 +0000 (20:57 +0000)]
Only query the colormap for GRAYSCALE and PSEUDOCOLOR visuals, and don't

Tue Feb  3 15:34:27 1998  Owen Taylor  <owt1@cornell.edu>

* gdk/gdkcolor.c (gdk_colormap_get_system): Only query
  the colormap for GRAYSCALE and PSEUDOCOLOR visuals,
  and don't ask for more than 256 colors in any case.

* gdk/gdkwindow.c (gdk_window_internal_destroy): Remove
  the input window information when we destroy the window,
  not when we are notified of it.

* gdk/gdkinputcommon.h (gdk_input_device_new): Work around
  bug in XFree86 3.3.1's handling of Wacom macro buttons.
  by assumming no device will report exactly 25 buttons.

* gdk/gdkinputcommon.h (gdk_input_common_other_event): Fill
  in string translation for generated key press events,
  do sanity checking on received key codes.

* gdk/gdkcc.c (gdk_color_context_new): Allocate enough
  room for a GdkColorContextPrivate, not just for a
  GdkColorContext.

26 years agodon't add the same menu to different menuitems/ optionmenus.
Tim Janik [Tue, 3 Feb 1998 15:14:35 +0000 (15:14 +0000)]
don't add the same menu to different menuitems/ optionmenus.

Tue Feb  3 15:09:55 1998  Tim Janik  <timj@gimp.org>

        * gtk/testgtk.c: don't add the same menu to different menuitems/
          optionmenus.

        * gtk/gtkmenuitem.h:
        * gtk/gtkmenuitem.c: new function gtk_menu_item_remove_submenu ro
          be consistent with optionmenu. use gtk_menu_attach_to_widget/
          gtk_menu_detach for setting/removing the submenu.
          invoke gtk_widget_destroy(submenu) in destructor to be consistent
          with other destructors.

        * gtk/gtkoptionmenu.h:
        * gtk/gtkoptionmenu.c: attach/detach to menu widget via
          gtk_menu_attach_to_widget/gtk_menu_detach.
          invoke gtk_widget_destroy(menu) in destructor to be consistent
          with other destructors.

26 years agoattach/detach to menu widget via
Tim Janik [Tue, 3 Feb 1998 14:13:05 +0000 (14:13 +0000)]
attach/detach to menu widget via

Tue Feb  3 15:09:55 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkoptionmenu.h:
        * gtk/gtkoptionmenu.c: attach/detach to menu widget via

        * gtk/gtkmenu.h:
        * gtk/gtkmenu.c: new functions gtk_menu_attach_to_widget
          and gtk_menu_detach that correspond to the action of
          gtk_widget_set_parent and gtk_widget_unparent.

        * gtk/widget.c: few fixups.

26 years agogtk/gtktable.c Fixed problem with division by zero in row/column-spanned
Owen Taylor [Tue, 3 Feb 1998 05:57:31 +0000 (05:57 +0000)]
gtk/gtktable.c Fixed problem with division by zero in row/column-spanned

Tue Feb  3 00:12:00 1998  Owen Taylor  <owt1@cornell.edu>
* gtk/gtktable.c
  Fixed problem with division by zero in row/column-spanned
  tables. Also removed a bunch of conditionals by making
  the observation that x/1 == x.

26 years agoFixed (?) problem with arithmentic exceptions in row/column
Owen Taylor [Tue, 3 Feb 1998 05:23:03 +0000 (05:23 +0000)]
Fixed (?) problem with arithmentic exceptions in row/column
spanned tables.

26 years agodunno, what cvs wants right now. i just told me these
Tim Janik [Mon, 2 Feb 1998 21:41:24 +0000 (21:41 +0000)]
dunno, what cvs wants right now. i just told me these
changes are already commited... ;(
-timj

26 years agochanged name to reflect that these macros in fact can't operate on
Tim Janik [Mon, 2 Feb 1998 21:07:18 +0000 (21:07 +0000)]
changed name to reflect that these macros in fact can't operate on

Mon Feb  2 04:15:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkwindow.c:
        * gtk/gtkwidget.c:
        * gtk/gtkmain.c:
        * gtk/gtkwidget.c:
        * gtk/gtkcontainer.c:
        * gtk/gtkprivate.h (GTK_PRIVATE_UNSET_FLAG) (GTK_PRIVATE_UNSET_FLAGS):
          changed name to reflect that these macros in fact can't operate on
          multiple flags.

26 years agofixed destroy handler, so it doesn't segfault with the new refcounting
Tim Janik [Mon, 2 Feb 1998 20:40:03 +0000 (20:40 +0000)]
fixed destroy handler, so it doesn't segfault with the new refcounting

Mon Feb  2 04:15:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtktoolbar.c: fixed destroy handler, so it doesn't
          segfault with the new refcounting scheme anymore.

        * gtk/gtkhandlebox.c:
        * gtk/gtkclist.c:
        * gtk/gtkentry.c:
        * gtk/gtkrange.c:
        * gtk/gtktext.c:
        * gtk/gtkviewport.c:
          enforced
          gdk_window_set_user_data (window, NULL);
          gdk_window_destroy (window);
          window = NULL;
          throughout the code.

26 years agofixed a bad, bad referencing bug that could caused unreferencing of
Tim Janik [Mon, 2 Feb 1998 18:44:28 +0000 (18:44 +0000)]
fixed a bad, bad referencing bug that could caused unreferencing of

Mon Feb  2 04:15:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkmain.c (gtk_propagate_event): fixed a bad, bad referencing
          bug that could caused unreferencing of finalized objects.

        * gtk/testgtk.c: destroy fileselection on "OK" (this triggered the
          above mentioned bug).

        * gtk/gtkwidget.h:
        * gtk/gtkwidget.c:
        * gtk/gtkobject.h:
        * gtk/gtkobject.c:
          implemented and object reference tracer (gtk_trace_referencing) which
          is activated if GTK_TRACE_OBJECTS is defined (currently per default).
          in gdb: set the static variable `gtk_trace_object' to point to the
          object that you want to have reference traced.

        * gtk/gtkfileselection.c: few cleanups.

26 years agocleanups with key name spaces.
Tim Janik [Mon, 2 Feb 1998 14:13:13 +0000 (14:13 +0000)]
cleanups with key name spaces.

Mon Feb  2 04:15:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkcolorsel.c:
        * gtk/gtkcombo.c:
        * gtk/gtkobject.c:
        * gtk/gtkselection.c:
        * gtk/gtkwidget.c:
        * gtk/gtkwindow.c:
          cleanups with key name spaces.

        * gtk/gtkcombo.c: fixed destruction bug with popwin.

26 years agowatch the next ChangeLog update...
Tim Janik [Mon, 2 Feb 1998 14:11:24 +0000 (14:11 +0000)]
watch the next ChangeLog update...
-timj

26 years agofixed a refresh bug with unrealization of no_window widgets.
Tim Janik [Mon, 2 Feb 1998 05:35:59 +0000 (05:35 +0000)]
fixed a refresh bug with unrealization of no_window widgets.
(dialog button in tesgtk)
-timj

26 years agoforgot gtkprivate.h as well
Tim Janik [Mon, 2 Feb 1998 04:56:42 +0000 (04:56 +0000)]
forgot gtkprivate.h as well
-timj

26 years agohehe, forgot the ChangeLog
Tim Janik [Mon, 2 Feb 1998 04:55:09 +0000 (04:55 +0000)]
hehe, forgot the ChangeLog
-timj

26 years agoGTK_RESIZE_NEEDED is a private flag now.
Tim Janik [Mon, 2 Feb 1998 04:54:25 +0000 (04:54 +0000)]
GTK_RESIZE_NEEDED is a private flag now.

Mon Feb  2 04:15:08 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkcontainer.h:
        * gtk/gtkcontainer.c: GTK_RESIZE_NEEDED is a private flag now.
          (gtk_container_register_toplevel): new function.
          (gtk_container_unregister_toplevel): new function.

        * gtk/gtkmain.c: GTK_LEAVE_PENDING is a private flag now.

        * gtk/gtkmenu.c: call gtk_container_register_toplevel in
          gtk_menu_class_init instead of this dirty gtk_widget_set_parent(,NULL)
          hack. new default handler gtk_menu_destroy for calling
          gtk_container_unregister_toplevel. removed GTK_ANCHORED, GTK_UNMAPPED.

        * gtk/gtkobject.h: macro cleanups, added GTK_DESTROYED flag.

        * gtk/gtkobject.c: only emit DESTROY signal if !GTK_OBJECT_DESTROYED
          (object).

        * gtk/gtkprivate.h: new file that will not be automatically included.
          it holds the private flags for GtkWidget along with it's SET/UNSET
          and examination macros.

        * gtk/gtkwidget.c: private flags: GTK_RESIZE_NEEDED, GTK_REDRAW_PENDING,
          GTK_RESIZE_PENDING, GTK_IN_REPARENT, GTK_USER_STYLE. GTK_ANCHORED is
          replaced by GTK_TOPLEVEL. added missing UNSET for GTK_IN_REPARENT.
          removed the gtk_widget_set_parent(, NULL) hack for toplevels.
          upon destroy free memory for widgets with GTK_WIDGET_HAS_SHAPE_MASK.

        * gtk/gtkwidget.h: split up the widget flags into a public and a private
          portion. added an extra field private_flags to GtkWidget without making
          it bigger by using an alignment gap of 16 bit. macro cleanups.

        * gtk/gtkwindow.c: removed GTK_ANCHORED. new function gtk_window_destroy
          for calling gtk_container_unregister_toplevel. removed the
          gtk_widget_set_parent(,NULL), call gtk_container_register_toplevel
          instead. remove GTK_UNMAPPED. GTK_RESIZE_NEEDED is private now.

        * gtk/gtksignal.c (gtk_signal_disconnect): removed a bug on
          removal that cut off the handler list -> living_objects == 0
          with testgtk. made some warnings more descriptive.
          new function gtk_signal_connect_object_while_alive, which
          will automatically destroy the connection once one of the objects
          is destroyed. didn't include this before removal of the above
          mentioned bug.

* reflected refcounting revolution in ChangeLog

26 years agomove enum GtkToolbarChildType before typedef to make gtktoolbar.h compile with c...
Radek Doulik [Sun, 1 Feb 1998 08:50:08 +0000 (08:50 +0000)]
move enum GtkToolbarChildType before typedef to make gtktoolbar.h compile with c++ (or egcs at least :)

26 years agoSome documentation of the internals.
Owen Taylor [Sun, 1 Feb 1998 08:22:51 +0000 (08:22 +0000)]
Some documentation of the internals.
                                        Owen

26 years agoAdded a new GTK_IN_REPARENT flag and revised gtk_widget_reparent so that
Owen Taylor [Sun, 1 Feb 1998 02:28:14 +0000 (02:28 +0000)]
Added a new GTK_IN_REPARENT flag and revised gtk_widget_reparent so that

Sat Jan 31 21:26:27 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkwidget.c: Added a new GTK_IN_REPARENT flag and revised
  gtk_widget_reparent so that the child would not be unrealized
  unnecessarily. Changed gtk_widget_unrealize () to recursively
  unrealize the children (since the child windows will be
  destroyed too.)

26 years agoI got sick of typing in the same commands over and over, so borrowed
Shawn Amundson [Sat, 31 Jan 1998 17:43:56 +0000 (17:43 +0000)]
I got sick of typing in the same commands over and over, so borrowed
autogen.sh from gnome cvs...  Also corrected some info in HACKING.

-Shawn

26 years agoAdded and autoconfigured in a new utility function g_strcasecmp
Manish Singh [Sat, 31 Jan 1998 08:15:57 +0000 (08:15 +0000)]
Added and autoconfigured in a new utility function g_strcasecmp
GtkCombo uses g_strcasecmp now
Quelled a warning about atexit undeclared in gtkobject

-Yosh

26 years agoBacked out change to close polygons since it caused problems for the
Owen Taylor [Sat, 31 Jan 1998 05:12:37 +0000 (05:12 +0000)]
Backed out change to close polygons since it caused problems for the

Sat Jan 31 00:13:33 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkstyle.c: Backed out change to close polygons
  since it caused problems for the Notebook.

26 years ago- Conformity to new reference counting schemes - Font and color
Owen Taylor [Sat, 31 Jan 1998 03:33:02 +0000 (03:33 +0000)]
- Conformity to new reference counting schemes - Font and color

Fri Jan 30 22:28:09 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkclist.{c,h}:
   - Conformity to new reference counting schemes
   - Font and color information is no-longer taken from
     the (possibly unitialized) style before the widget
     is realized, but is obtained when needed. (Fixes
     problems with non-default styles)
   - white_gc => base_gc where appropriate
   - Handle graphics expose events correctly (or as
     correctly as anywhere else) so that scrolling when
     overlapped works.

26 years agoIn gtk_widget_init, set the flags _on_top_of_ those for
Owen Taylor [Sat, 31 Jan 1998 02:28:43 +0000 (02:28 +0000)]
In gtk_widget_init, set the flags _on_top_of_ those for
GTK_OBJECT, to avoid clearing the FLOATING flag.
living_objects == 0!

26 years agofixed round-off error in calculation of tables with a lot of cells
George Lebl [Sat, 31 Jan 1998 01:43:47 +0000 (01:43 +0000)]
fixed round-off error in calculation of tables with a lot of cells

Fri Jan 30 17:41:06 1998  George Lebl  <jirka@5z.com>

        * gtk/gtktable.c: fixed round-off error in calculation
          of tables with a lot of cells

26 years agored/green/blue parameters are now expected to be in [0, 65535], to be
Federico Mena [Sat, 31 Jan 1998 00:09:58 +0000 (00:09 +0000)]
red/green/blue parameters are now expected to be in [0, 65535], to be

Fri Jan 30 08:51:16 1998  Federico Mena  <federico@bananoid.nuclecu.unam.mx>

* gdk/gdkcc.c (gdk_color_context_get_pixel): red/green/blue
  parameters are now expected to be in [0, 65535], to be consistent
  with the rest of Gdk.
(gdk_color_context_get_pixels): Made it use 16-bit color values as
  well.  Fixed mdist=1000000 buglet (it should start with at least
  0x1000000).
(gdk_color_context_get_pixels_incremental): Same as
gdk_color_context_get_pixels().

26 years agohm, initital refcount revolution commit ;)
Tim Janik [Fri, 30 Jan 1998 23:47:09 +0000 (23:47 +0000)]
hm, initital refcount revolution commit ;)
still some gnits left, but keep working on it ;)
-timj

26 years agomake toolbar Child structure accesible from apps
Radek Doulik [Fri, 30 Jan 1998 10:21:37 +0000 (10:21 +0000)]
make toolbar Child structure accesible from apps

26 years agoClose the polygon if it isn't already. (To match gtk_draw_polygon)
Owen Taylor [Fri, 30 Jan 1998 03:58:47 +0000 (03:58 +0000)]
Close the polygon if it isn't already. (To match gtk_draw_polygon)

Thu Jan 29 22:57:39 1998  Owen Taylor  <owt1@cornell.edu>

* gtk/gtkstyle.c (gtk_default_draw_polygon): Close
  the polygon if it isn't already. (To match gtk_draw_polygon)
  Simplified logic. (Appearance could probably be
  improved for objects with gradual curves by adding in
  some intermediate edge coloration)

26 years agodiscovered and documented a new bug ;)
Tim Janik [Fri, 30 Jan 1998 03:51:43 +0000 (03:51 +0000)]
discovered and documented a new bug ;)
-timj

26 years agosome bugs worked out.
Tim Janik [Fri, 30 Jan 1998 01:34:19 +0000 (01:34 +0000)]
some bugs worked out.
comment cleanups
-timj

26 years agonew function to gather information about a certain signal.
Tim Janik [Thu, 29 Jan 1998 20:46:05 +0000 (20:46 +0000)]
new function to gather information about a certain signal.

Thu Jan 29 21:36:14 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtksignal.c (gtk_signal_query): new function to gather
          information about a certain signal.

        * gtk/gtksignal.c (gtk_signal_newv): new function similar to
          gtk_signal_new().

        * gtk/gtksignal.c (gtk_signal_real_emit): check for function_offset
          == 0.

        * gtk/gtksignal.c (gtk_signal_connect_by_type): perform a signal
          id lookup on the parent as well.

        * gtk/gtkobject.c (gtk_object_class_add_user_signal): new
          function for implementation of user defined signals.
        * gtk/gtkobject.c (gtk_object_class_add_signals): free old
          signal id array.

        * gtk/gtkobject.h: this holds the typedefs for GtkSignalFunc and
          GtkSignalMarshaller now, because they are used for
          gtk_object_class_add_user_signal.

        * gtk/gtktypeutils.c (gtk_type_class_init): reset object_class->signals
          and object_class->nsignals for new object classes.

26 years agonew user signal, and possibility to query signal information.
Tim Janik [Thu, 29 Jan 1998 20:44:14 +0000 (20:44 +0000)]
new user signal, and possibility to query signal information.
-timj

26 years agothis goes to *all* widget writers,
Tim Janik [Thu, 29 Jan 1998 13:47:47 +0000 (13:47 +0000)]
this goes to *all* widget writers,
you have to follow the gtk+ nameing scheme, otherwise bad things
will happen (in e.g. signal/argument/type lookups).
-timj

26 years agoremoved a FIXME that i had forgotten earlier.
Tim Janik [Thu, 29 Jan 1998 08:44:55 +0000 (08:44 +0000)]
removed a FIXME that i had forgotten earlier.

26 years agoThe list sort functions compared things backwards. Fixed.
Manish Singh [Thu, 29 Jan 1998 08:00:28 +0000 (08:00 +0000)]
The list sort functions compared things backwards. Fixed.
The list iterator macros now check for NULL args

-Yosh

26 years agoSmall fix to my previous GtkTable commit -miguel
Arturo Espinosa [Wed, 28 Jan 1998 05:55:54 +0000 (05:55 +0000)]
Small fix to my previous GtkTable commit -miguel

26 years agoOops, forgot a clean-up - Federico
Arturo Espinosa [Wed, 28 Jan 1998 03:51:48 +0000 (03:51 +0000)]
Oops, forgot a clean-up - Federico

26 years agoThe auto_shrink policy of the floating window is now set to TRUE. This
Federico Mena [Wed, 28 Jan 1998 03:41:55 +0000 (03:41 +0000)]
The auto_shrink policy of the floating window is now set to TRUE. This

Tue Jan 27 15:52:48 1998  Federico Mena  <federico@bananoid.nuclecu.unam.mx>

* gtk/gtkhandlebox.c (gtk_handle_box_realize): The auto_shrink
policy of the floating window is now set to TRUE.  This fixes the
problem of the floating window being too big when the handlebox
child is small.
(gtk_handle_box_motion): Now we use GDK_POINTER_MOTION_HINT_MASK
to improve movement.

26 years agoTables autoresize now -miguel
Arturo Espinosa [Tue, 27 Jan 1998 23:40:51 +0000 (23:40 +0000)]
Tables autoresize now -miguel

26 years agoMore changes from Tony Gale.
Shawn Amundson [Tue, 27 Jan 1998 20:51:36 +0000 (20:51 +0000)]
More changes from Tony Gale.

26 years agog_string_prepend* had interchanged src and dest params. Fixed.
Manish Singh [Tue, 27 Jan 1998 17:50:07 +0000 (17:50 +0000)]
g_string_prepend* had interchanged src and dest params. Fixed.

-Yosh

26 years agoFixed a really lame error where g_slist_insert didn't hook the data
Manish Singh [Tue, 27 Jan 1998 09:42:41 +0000 (09:42 +0000)]
Fixed a really lame error where g_slist_insert didn't hook the data
in! Reworked the routine a bit too.

-Yosh

26 years agoShows the widgets properly. -Jay
Jay Painter [Tue, 27 Jan 1998 06:39:56 +0000 (06:39 +0000)]
Shows the widgets properly. -Jay

26 years agoFixed widget sensitivity for the treeitem - please review at leisure
Elliot Lee [Tue, 27 Jan 1998 03:57:59 +0000 (03:57 +0000)]
Fixed widget sensitivity for the treeitem - please review at leisure

26 years agoadditions/changes from new FAQ maintainers Nathan Froyd and Tony Gale.
CST 1998 Shawn T. Amundson [Mon, 26 Jan 1998 23:11:51 +0000 (23:11 +0000)]
additions/changes from new FAQ maintainers Nathan Froyd and Tony Gale.

Mon Jan 26 17:12:12 CST 1998 Shawn T. Amundson <amundson@gimp.org>

        * docs/gtkfaq.sgml: additions/changes from new FAQ maintainers
          Nathan Froyd and Tony Gale.

26 years agoadditions/changes from Gregory A. McLean
CST 1998 Shawn T. Amundson [Mon, 26 Jan 1998 22:57:10 +0000 (22:57 +0000)]
additions/changes from Gregory A. McLean

Mon Jan 26 16:54:02 CST 1998 Shawn T. Amundson <amundson@gimp.org>

        * docs/gtk.texi: additions/changes from Gregory A. McLean

26 years agoadded a bunch of checks for window != NULL and private->destroyed. (don't
Tim Janik [Mon, 26 Jan 1998 01:20:14 +0000 (01:20 +0000)]
added a bunch of checks for window != NULL and private->destroyed. (don't

Mon Jan 26 02:15:29 1998  Tim Janik  <timj@gimp.org>

        * gdk/gdkdraw.c:
        * gdk/gdkgc.c:
        * gdk/gdkimage.c:
        * gdk/gdkinput.c:
        * gdk/gdkpixmap.c:
        * gdk/gdkproperty.c:
        * gdk/gdkselection.c:
        * gdk/gdkwindow.c: added a bunch of checks for window != NULL and
          private->destroyed.
(don't trust this cvs commit message, i didn't modify that many files!)

26 years agoFixed a resizing bug with CList where you could loose the end of the last
Jay Painter [Mon, 26 Jan 1998 00:07:40 +0000 (00:07 +0000)]
Fixed a resizing bug with CList where you could loose the end of the last
column; I also fixed a small cosmetic problem with the fileselect. -Jay

26 years agoadded new widget flag GTK_LEAVE_PENDING. if a widget has GTK_LEAVE_PENDING
Tim Janik [Sun, 25 Jan 1998 18:45:56 +0000 (18:45 +0000)]
added new widget flag GTK_LEAVE_PENDING. if a widget has GTK_LEAVE_PENDING

Sun Jan 25 19:15:32 1998  Tim Janik  <timj@gimp.org>

        * gtk/gtkwidget.h (enum): added new widget flag GTK_LEAVE_PENDING.
        * gtk/gtkmain.c (gtk_main_iteration_do): if a widget has
          GTK_LEAVE_PENDING set, send it its LEAVE_NOTIFY event, regardless
          of a grab or sensitivity.
          changed the compression code for enter/leave notify events to
          free *both* compressed event, and removed an unneccessary call
          to g_list_remove.

        * ChangeLog: finally catched up with the recent changes on my part.

26 years agoforgot a printf() line ;)
Tim Janik [Sat, 24 Jan 1998 20:03:58 +0000 (20:03 +0000)]
forgot a printf() line ;)
-timj

26 years agogtk_object_query_args() now returns the args in the corret order.
Tim Janik [Sat, 24 Jan 1998 20:02:19 +0000 (20:02 +0000)]
gtk_object_query_args() now returns the args in the corret order.
-timj

26 years agoChanged static function names/reorganized to make more sense. -Jay
Jay Painter [Sat, 24 Jan 1998 19:14:42 +0000 (19:14 +0000)]
Changed static function names/reorganized to make more sense. -Jay

26 years agoadded BUGS. -timj
Tim Janik [Sat, 24 Jan 1998 18:45:34 +0000 (18:45 +0000)]
added BUGS. -timj

added BUGS.
-timj

26 years agonew element n_args in GtkObjectClass.
Tim Janik [Fri, 23 Jan 1998 10:38:29 +0000 (10:38 +0000)]
new element n_args in GtkObjectClass.
-timj

26 years agoSwitched the file to GNU indentation, for consistency.
Federico Mena [Fri, 23 Jan 1998 06:26:39 +0000 (06:26 +0000)]
Switched the file to GNU indentation, for consistency.

Thu Jan 22 18:58:44 1998  Federico Mena  <federico@bananoid.nuclecu.unam.mx>

* gdk/gdkcc.c: Switched the file to GNU indentation, for consistency.

26 years agoPicky, picky... fix two lines with wrong indentation :-)
Federico Mena [Fri, 23 Jan 1998 06:05:39 +0000 (06:05 +0000)]
Picky, picky... fix two lines with wrong indentation :-)

Thu Jan 22 18:58:44 1998  Federico Mena  <federico@bananoid.nuclecu.unam.mx>

* gdk/gdktypes.h (struct _GdkColorContext): Picky, picky... fix
two lines with wrong indentation :-)

26 years agogtk/Makefile.am: Fixed the called to runelisp to use $(SHELL).
scott [Thu, 22 Jan 1998 07:34:13 +0000 (07:34 +0000)]
gtk/Makefile.am: Fixed the called to runelisp to use $(SHELL).
(Cannot rely on the script having execute permissions.)
--sg