This patch adds the GtkScrollablePolicy type property to GtkScrollable
and implements it in all subclasses. GtkScrolledWindow observes this
property to make a good guess about when to show/hide scrollbars for
height-for-width content.
Most scrollable children do not do height-for-width *yet* but
most certainly will (toolpalette, treeview, iconview, textview
widgets all TODO), for scrollable widgets that do have a minimum
and natural size, it's important for them to observe the state
of this property in order to properly drive the scroll adjustments
according to the desired GtkScrollablePolicy. This patch makes
GtkViewport do this.
Patch also adds tests/testscrolledwindow.c to display the effects
of this property.
The scrollable patch removed set_scroll_adjustments (NULL, NULL) from
gtk_tree_view_init() which ensured the treeview would have adjustments
at all times, this patch adds set_v/hadjustment (NULL) to _init to ensure
the same thing (without it, there are crashes when trying to access the
adjustments notably from set_headers_visible() in gedit).
I.e. Since we are now calling get_preferred_width() to ensure a good 'for_size'
for get_height_for_width() we need to avoid warning about this internal expected
recursion.
These properties use GdkRGBA to render the cell renderer's content.
Note that Pango attributes are used to render the foreground color,
so the alpha value is currently ignored.
It is just too annoying to have to implement these properties in
every scrollable. Instead, we now have ::min-content-height/width
in GtkScrolledWindow.
We also add GtkScrollablePolicy to determine how to size the
scrollable content.
Since the ::changed implementation of GtkRecentManager implies a
synchronous write operation, when we receive multiple requests to emit a
::changed signal we might end up blocking.
This change coalesces multiple ::changed emission requests using the
following sequence:
• the first request will install a timeout in 250 ms, which will
emit the ::changed signal
• each further request while the timeout has not been emitted
will increase a counter
‣ if the counter reaches 250 before the timeout has been
emitted, then the RecentManager will remove the timeout
source and force a signal emission and reset the counter
This sequence should guarantee that frequent ::changed emission requests
are coalesced, and also guarantee that we don't let them dangle for too
long.
https://bugzilla.gnome.org/show_bug.cgi?id=616997
alignment/margin vfuncs adjust_size_request/allocation
Now get_height_for_width() will internally update the for_width
before passing it to the real height_for_width() vfunc, allowing
margins and extra space for alignments to be stripped, thus requesting
sufficient height for greater than natural widths (and also accounting
for margins properly). Test case adjusted in testadjustsize to ensure
proper behavior.
Create a new adjustment if there was *no* adjustment passed in, not
the other way around. Also remove unused function
gtk_tool_palette_set_adjustment().
The GtkScrollable interface provides "hadjustment" and "vadjustment"
properties that are used by GtkScrolledWindow. It replaces
the ::set_scroll_adjustment signal. The scrollable interface
also has ::min-display-width/height properties that can be
used to control the minimally visible part inside a scrolled window.
Use input-only windows for capturing input, and just draw onto
the parent window. This allows themes to make entries with rounded
corners. Patch by Boram Park.
https://bugzilla.gnome.org/show_bug.cgi?id=632736
Remove the check for GTK_TREE_MODEL_LIST_ONLY and make the view
display only the first level of the tree, ignoring all insert, remove
etc. events on the model's branches. This enables generic application
code which always produces GtkTreeStores (also for flat lists) to
display grids of these flat trees.
We lookup a widget to call a function on GtkWidget that ends up just
performing an operation directly on the GdkWindow again anyway
(beeping). There's already fallback code at the point of the call that
does exactly that anyway, so just do that all of the time.
- add slots for damage-event, move-focus and keynav-failed
- reorder signals a bit so related stuff is grouped together
- some indentation fixes in the GtkWidgetClass
- remove the move-focus compat hack from GtkTextView
- turn the move-focus compat hack in GtkWindow into properly
implementing GtkWidget::move-focus()
The gtkprivate.h header contains GtkWidget-specific private symbols that
are not useful except in a handful of cases. Basically everything
includes gtkprivate.h for the GTK_PARAM_* macros.
https://bugzilla.gnome.org/show_bug.cgi?id=632539
Some GtkSettings property are registered by other classes. This leads
to the "interesting" issue that setting GtkSettings:gtk-button-images
requires that the GtkButton class is referenced first - or that a
GtkButton is created.
https://bugzilla.gnome.org/show_bug.cgi?id=632538
This path would only do anything where widget->window was not located at
widget->allocation.x/y but at a different position. This should never
happen in the real world. But if it does, GTK3 will break for you.
Better fix your widget.
gtk_icon_view_layout(), which can resize the bin window and thus
underlying surfaces, could be called from gtk_icon_view_draw(). So
we have to make sure that layouting is done before the draw handler is
called. We use the same construction as in tree view here, the layout
idle handler is installed with a higher priority than
GDK_PRIORITY_REDRAW and we wrap the call to gdk_window_process_updates()
in a new function that will flush the layout queue beforehand.
This patch adds a GtkComboBox:popup-fixed-width to decide if the
popup's width should be a fixed width matching the combo's allocated width.
The patch includes a kindof hack to work around treeviews currently
not supporting height-for-width geometry (for list-mode only), this
hack can be safely removed once treeviews start reporting natural
widths properly.
After the filechooser got converted to use gsettings, it can no
longer be used uninstalled, breaking make distcheck here.
Also disable the treeview-scrolling tests, since they are failing
here.
Parse options job-sheets, job-hold-until and sides correctly.
Add get_lpoption_name() for translation of lpoption names to
gtk option names. Usable for options which values don't need
conversion (e.g. number-up, number-up-layout, job-billing
and job-priority).
Rename array option_names to ppd_option_names to reflect its
purpose better. Rename get_option_name() to get_ppd_option_name()
because of the same reason.
This is a derived class of GtkComboBox to replace the
gtk_combo_box_*_text() convenience API.
Fixes https://bugzilla.gnome.org/show_bug.cgi?id=612396
Signed-off-by: Javier Jardón <jjardon@gnome.org>
We cannot use the GtkButton::clicked signal to override the default
behavior of GtkLinkButton (i.e. call gtk_show_uri()), because
GtkButton::clicked is registered as a RUN_FIRST signal, which obviously
prevents any other signal handler connected to it to stop the
propagation before the class handler has a chance to run.
For this reason we can add a GtkLinkButton::activate-link signal, which
will be emitted by the default GtkButton::clicked signal handler; the
::activate-link signal has a boolean return value, which allows simpler
code for stopping the propagation to the next signal handler.
https://bugzilla.gnome.org/show_bug.cgi?id=632150
Removed portion of scrolled-window that observes user-set
size request data (aux_info->width/height) on the child directly
in order to derive it's minimum possible size... if the scrolled
window has (auto/always) scrollbars in a said orientation; only
request enough space for the scrollbars (bug 631976).
The idea is that it is way more common to want to manipulate the
actual list, rather than the headers. Once you Tab into the treeview
(and the list part gets the focus), you can use Shift-Tab to focus
the headers.
This means that some hysteresis is added to the focus chain, but
it makes the treeview a lot more convenient to focus with the keyboard.
Signed-off-by: Federico Mena Quintero <federico@novell.com>
Because validate_visible_area() can modify the window size (and thus
the underlying surface), it should not be called from within the draw
method. Given that the presize handler is run with a higher priority
than redraw, and the presize handler will validate the visible area,
there should not be cases wherein the draw method is called and
validate_visible_area() has not been run yet.
However, one such a case was gdk_window_process_updates(), which would
trigger the draw method at some point. We now work around this
by factoring this in a new gtk_tree_view_bin_process_updates() function
that will run the presize handler first if needed.
Note: for other platforms, it might still be the case that the draw
method is called and validate_visible_area() has not been run yet.
(For example the Mac backend calls gdk_window_process_updates() from
the drawRect method, and the redraw-in-idle handling thus works
differently). This does not seem to be a problem now, if
it will be in the future we need to take care of that then.
This test is constructed in such a way that it breaks when you
(currently!) do not trigger a size-request from validate_visible_area.
Especially row expansion appears to have a need for this currently.
GtkComboBox now sports a construct-only "has-entry" property which
decides if it uses a GtkEntry to allow additional user input. Also
it has a new "entry-text-column" to fetch strings for the entry
from the model.
This patch deprecates the GtkComboBoxEntry and updates the rest of GTK+
to use the new semantics on GtkComboBox instead.
GtkComboBoxEntry will be removed altogether before GTK+ 3, in a
later commit.
We expand a tab if either tab-expand is set, or the generic widget
expand property in the correct direction is set. And we do not
propagate expand flags from tab labels to the notebook, only
the expand flags from the pages.
We expand a child when it either has the table-specific expand flag
or the generic expand property set. Override compute_expand so that
it also takes the table-specific expand flags of children into
account.
https://bugzilla.gnome.org/show_bug.cgi?id=628902
This consists of:
* expand a child if either child->expand || gtk_widget_get_expand(child)
* override compute_expand so that child->expand will cause us to
return TRUE for gtk_widget_get_expand()
https://bugzilla.gnome.org/show_bug.cgi?id=628902
GtkWidget now has flags for horizontal and vertical expand, and
a compute_expand() method. compute_expand() is used by containers
to set a default expand flag. (If a widget has expand set explicitly,
it always overrides the results of compute_expand.)
GtkContainer has a default implementation of compute_expand which
simply walks over all child widgets and sets expand=TRUE
if any child is expanding.
The overall effect is that you only have to set expand on
leaf nodes in the widget tree, while previously you had to
set expand all the way up the tree as you packed every
container. Also, now containers need not have their own child
properties for expand.
For old containers which do have "expand" child properties,
they should override compute_expand and include the child
properties in whether the container is expand=TRUE.
Also, the old container should use
"child_prop_expand || gtk_widget_compute_expand()" everywhere
it previously used simply "child_prop_expand"
https://bugzilla.gnome.org/show_bug.cgi?id=628902
Safe space in the private struct, always set the cursor when we
update visibility, since the cursor depends on resizability conditions
too. And don't set the shape needlessly.
This was caused by get_drag_edge() returning -1 to mean 'no drag',
but callers didn't really pay attention. Change get_drag_edge()
to return a boolean instead.
Call gtk_tree_view_set_cursor() in addition to
gtk_tree_selection_select_iter() when selecting the file in
show_and_select_files() so the preview update machinery gets
triggered.
(cherry picked from commit 795c8070db)
This patch makes the scrolled window reconsider allocating the child
the full width or height (depending on the child's request mode) without
a scrollbar. For instance when the child is height-for-width; the child
will first be tested if the content's height for full allocated width
(without a vscrollbar) will allow the contents height for that width
to fit the allocated height.
Patch is a simplified version of code inspected in st-scroll-view.c.
Note that this patch assumes children will begin to scroll only after
reaching their minimum size; adding a property to the future
GtkScrollableIface to decide whether to scroll-to-minimum or scroll-to-natural
will effect this code (it should then reconsider whether the child
will scroll below the natural size instead of the minimum).
Patch addresses bug 629778.
If you set a geometry widget via gtk_window_set_geometry_hints() it
becomes very hard to compute appropriate toplevel sizes in pixels
to make the window a particular size. Synthesizing strings and passing
them to gtk_window_parse_geometry() is possible, but to avoid
avoid such ugliness, add functions:
gtk_window_set_default_geometry()
gtk_window_resize_to_geometry()
That act like gtk_window_set_default_size() and
gtk_window_resize() but are in terms of the resize increments of the
geometry widget.
https://bugzilla.gnome.org/show_bug.cgi?id=631796
gtk_window_parse_geometry() gets the size of the window in order
to interpret the position of the window; calling it before, say,
calling gtk_widget_show_all() on a window is a subtle trap, so
add a warning in the case we can easily detect.
https://bugzilla.gnome.org/show_bug.cgi?id=631794
The geometry widget feature of gtk_window_set_geometry_hints() has
never really worked right because the calculation that GTK+ did to
compute the base size of the window only worked when the geometry
widget had a larger minimum size than anything else in the window.
Setup:
* Move the GtkSizeGroup private functions to a new private header
gtksizegroup-private.h
* Add the possibilty to pass flags to _gtk_size_group_queue_resize(),
with the flag GTK_QUEUE_RESIZE_INVALIDATE_ONLY to suppress adding
the widget's toplevel to the resize queue.
* _gtk_container_resize_invalidate() is added to implement that feature
* _gtk_widget_override_size_request()/_gtk_widget_restore_size_request()
allow temporarily forcing a large minimum size on the geometry
widget without creating resize loops.
GtkWindow:
* Compute the extra width/height around the geometry widget
correctly; print a warning if the computation fails.
* Always make the minimum size at least the natural minimum
size of the toplevel; GTK+ now fails badly with underallocation.
* Always set the base size hint; we were failing to set it
properly when the specified minimum size was overriden, but
it's harmless to always set it.
Tests:
* New test 'testgeometry' that replaces the 'gridded geometry' test
from testgtk. The new test is roughly similar but creates a bunch
of windows showing different possibilities.
* The testgtk test is removed. No need to have both.
https://bugzilla.gnome.org/show_bug.cgi?id=68668
1,936 (112 direct, 1,824 indirect) bytes in 4 blocks are definitely lost in loss record 13,453 of 13,673
at 0x4005BDC: malloc (vg_replace_malloc.c:195)
by 0x69615A1: cairo_region_create (cairo-region.c:196)
by 0x683BFBF: gdk_cairo_region_create_from_surface (gdkcairo.c:455)
by 0x670C882: set_grip_shape (gtkwindow.c:5020)
by 0x670CFBF: resize_grip_create_window (gtkwindow.c:5271)
by 0x670C466: gtk_window_realize (gtkwindow.c:4902)
Signed-off-by: Benjamin Otte <otte@redhat.com>
We need to be a little more careful when determining the overlap
between the new allocation and the grip area. This was causing
vertical scrollbars in evince to overlap with the grip.
This is safer, since GtkWindow subclasses may override realize
and end up with has_resize_grip == TRUE, but grip_window still
being NULL - as is the case with GtkPlug.
We need to position the grip window in size_allocate after all,
otherwise the grip ends up in the wrong place when the window
is realized early.
Also, avoid shadowing a variable.
This was another case of making sure size-request was called
before other code was doing get_child_requisition(), now caching
is done via the same api so no need for that line.
Old code was probably doing size-request on all children initially,
and then calling get_child_requisition() in other passes, now these
are cached after the first request anyway.
