patches that didn't get through

[Tim Janik <Tim Janik Hamburg Netsurf DE>]

gtk-timj-970903-0.patch:	(against gtk+970828.tar.gz)

the contents of this patch
(ftp.gimp.org/pub/gtk/patches/gtk-timj-970903-0.patch)
was originaly sent to gtk-list in 4 parts,
but didn't get through.
so, i'm just quoting my patch anouncements here.

	Tim Janik <timj@psynet.net>



> Date: Sun, 31 Aug 1997 10:44:09 +0200 (CEST)
> Subject: [patch] gtk?paned.c and gtkmenuitem.c
hi,

there is a small bug in gtkmenuitem.c, a new submenu
can be set if the old one is invisible and not if it
is visible.

the diff also includes a patch against gtkhpaned.c
and gtkvpaned.c that prevents a segmentation fault if
the paned windows don't have children, because my last
mail to gtk-list about this didn't get trough.

this patch is against gtk+970828.tar.gz

---
ciaoTJ

> Date: Sun, 31 Aug 1997 16:28:58 +0200 (CEST)
> Subject: [patch] gtkmisc.c, gtkframe.c, gtk*button.c
hi,

this patch fixes several bugs as described below.
i've tested it with various draw_indicator, border_with, child
widgets with and without window combinations and am quite sure that it
works correctly.
because the button behaviour with border_width>0 was so horribly
broken, i'd suggest everyone who is going to alter the changes
this patch makes, to take a look at the window created by
gubi with the following lines - *before* and *after* appliyng the patch.

GB_WIDGET_WINDOW "" {
  resize_policy     6
  GB_WIDGET_FRAME "" {
    border_width      20
    label_xalign      1.
    GB_WIDGET_V_BOX "" {
      GB_WIDGET_BUTTON "" {
        border_width      20
      }
      GB_WIDGET_TOGGLE_BUTTON "" {
        border_width      20
      }
      GB_WIDGET_CHECK_BUTTON "" {
        border_width      30
      }
      GB_WIDGET_RADIO_BUTTON "" {
        border_width      20
        draw_indicator    0
      }
    }
  }
}

the following will be fixed by this patch:
the frame was not shown, it walked (with ongrowing border-with more and more
to the right side) out of the frame->window.
changing the alignment of a GtkMisc widget as well as the alignment and/or
text string of a frame label will now clear the text in the old position.
with descendants of GtkButton (GtkToggle/Check/RadioButton) and
GTK_CONTAINER()->border_with>0:
the shadow will not be drawn within the childs allocated area anymore.
if draw_indicator is TRUE (default) labels inside a button will draw
themselves again.
the check/radio button indicator will move to the right with growing
border_width.
also some strange shadow drawing behaviour with GTK_CAN_DEFAULT is
fixed for toggle buttons.


---
ciaoTJ

> Date: Sun, 31 Aug 1997 22:06:27 +0200 (CEST)
> Subject: Re: [patch] gtktooltips.c and gtkviewport.c
hi again,

yes i know it's the third patch for this day and the
only thing i should do is go into bed... i will RSN ;)

with this patch i got the last quirks out of gtk+ that showed up
in my last version of gubi, so i promise not to send out anymore
patches this day ;)))

the GtkTooltips used to unmap a GtkTooltipdata if the "destroy"
signal of a widget is emitted. this needs to be removed wich is done
by gtk_tooltips_widget_remove() now. otherwise we get strange
segfaults.

in a GtkViewport the GdkWindow didn't get realized (created) when
the viewport is child of a scrolled window that is in a page of a
GtkNotebook that is not the top most one.

please excuse any spelling errors, i'm quite tired, but don't worry
the patch is correct ;)

---
ciaoTJ

> Date: Wed, 3 Sep 1997 20:02:43 +0200 (CEST)
> Subject: Re: [patch] gtkaccelerator.c
hi,

this patch prevents gtk+ from segfaulting if accelerators are used
with values >127 e.g. the german 'ö' (oe).
the index for GtkAcceleratorTable.entries[] needs to be in the
range 0-255 and therefore an unsigned char is needed as index type.


---
ciaoTJ