]> Pileus Git - ~andy/gtk/log
~andy/gtk
12 years agoMerge branch 'win32-theme2'
Alexander Larsson [Mon, 21 Nov 2011 12:49:44 +0000 (13:49 +0100)]
Merge branch 'win32-theme2'

12 years ago[icons] text-x-generic isn't a generated icon
John Ralls [Mon, 21 Nov 2011 00:19:59 +0000 (16:19 -0800)]
[icons] text-x-generic isn't a generated icon

Calling it one causes maintainer-clean to delete it.

12 years agogtk/gtkapplication.c: Fix typo
Javier Jardón [Sat, 19 Nov 2011 16:30:43 +0000 (16:30 +0000)]
gtk/gtkapplication.c: Fix typo

12 years agoframe: make sure to allocate the CSS border width
Cosimo Cecchi [Fri, 18 Nov 2011 16:58:56 +0000 (11:58 -0500)]
frame: make sure to allocate the CSS border width

Similar to GtkNotebook, GtkFrame was only allocating space for the
padding width, and not the border.

This could be seen by just running tests/testframe. With a theme that
renders frame borders, setting xthickness = 0 in the test draws the
button border over the frame border, which is wrong.

https://bugzilla.gnome.org/show_bug.cgi?id=664342

12 years agoframe: add GTK_STYLE_CLASS_FRAME in _init()
Cosimo Cecchi [Fri, 18 Nov 2011 16:42:01 +0000 (11:42 -0500)]
frame: add GTK_STYLE_CLASS_FRAME in _init()

Instead of adding it every time we use the GtkStyleContext, just add it
in _init().

https://bugzilla.gnome.org/show_bug.cgi?id=664342

12 years agonotebook: make sure to allocate the CSS border width
Cosimo Cecchi [Fri, 18 Nov 2011 16:35:30 +0000 (11:35 -0500)]
notebook: make sure to allocate the CSS border width

Instead of taking only the CSS padding into account when allocating the
notebook children, also allocate the border width.

https://bugzilla.gnome.org/show_bug.cgi?id=664342

12 years agoBug 653676 - Expand/Collapse doesn't respond after one click
Rui Matos [Sun, 20 Nov 2011 20:00:08 +0000 (21:00 +0100)]
Bug 653676 - Expand/Collapse doesn't respond after one click

Ignore leave-notify-event when mode is GDK_CROSSING_GTK_[UN]GRAB.
Just [un]grabbing shouldn't cause us to [un]prelight the current arrow
and node.

12 years agoDon't call scroll to cell if the tree view is empty
Kristian Rietveld [Sun, 20 Nov 2011 17:54:20 +0000 (18:54 +0100)]
Don't call scroll to cell if the tree view is empty

Account for the case when gtk_drag_finish() didn't actually add a row
(can happen when dragging from an empty tree view to itself ...).

12 years agoBug 660554 - gtk_tree_view_drag_begin: assertion `path != NULL' failed
Kristian Rietveld [Sun, 20 Nov 2011 17:51:14 +0000 (18:51 +0100)]
Bug 660554 - gtk_tree_view_drag_begin: assertion `path != NULL' failed

Turned assertion into silent return.

This assertion is only hit when dragging from an empty tree view.  In
this case, gtk_tree_view_begin_drag() is triggered from gtkdnd.c and not
from gtk_tree_view_maybe_begin_dragging_row().  We actually want to
cancel the drag at this point, but that is not possible with the GTK+
API as far as I can see.

The alternative is to not allowing the drag to start.  This could be
done by simply unsetting the tree view as drag source when it is empty
and setting it as drag source again when rows are added.  I didn't
choose to go with this for now, since this will likely break third party
code.

12 years agoquartz: fix coding style in quartz-version of gtk_drag_begin_internal
Kristian Rietveld [Sun, 20 Nov 2011 17:27:28 +0000 (18:27 +0100)]
quartz: fix coding style in quartz-version of gtk_drag_begin_internal

12 years agoquartz: do not call gdk_drag_begin() twice
Kristian Rietveld [Sun, 20 Nov 2011 17:24:56 +0000 (18:24 +0100)]
quartz: do not call gdk_drag_begin() twice

This makes drag and drop work again on Quartz.

12 years ago[Bug 664238] GTK apps crash when dragging something
John Ralls [Sat, 19 Nov 2011 19:33:28 +0000 (11:33 -0800)]
[Bug 664238] GTK apps crash when dragging something

Corrects a bad condition in a test in 085b98f4

12 years agoUpdated Spanish translation
Jorge González [Sun, 20 Nov 2011 09:40:15 +0000 (10:40 +0100)]
Updated Spanish translation

12 years agoAdapt iconview a11y test output
Matthias Clasen [Sun, 20 Nov 2011 00:53:43 +0000 (19:53 -0500)]
Adapt iconview a11y test output

12 years agoImprove AtkText implementations
Matthias Clasen [Sun, 20 Nov 2011 00:28:48 +0000 (19:28 -0500)]
Improve AtkText implementations

There was some regressions from the recent display line fix;
while fixing it improve the test coverage and make GtkEntry
and GtkTextView return identical results.

12 years agoRemove leftover debug spew
Matthias Clasen [Sat, 19 Nov 2011 23:13:42 +0000 (18:13 -0500)]
Remove leftover debug spew

12 years agoAdd a missing static
Matthias Clasen [Sat, 19 Nov 2011 23:08:18 +0000 (18:08 -0500)]
Add a missing static

12 years agoFix the build
Matthias Clasen [Sat, 19 Nov 2011 22:57:08 +0000 (17:57 -0500)]
Fix the build

12 years agoUpdates
Matthias Clasen [Sat, 19 Nov 2011 21:15:46 +0000 (16:15 -0500)]
Updates

12 years agoDrop the Beagle search engine
Matthias Clasen [Sat, 19 Nov 2011 20:32:12 +0000 (15:32 -0500)]
Drop the Beagle search engine

It is not actively maintained.

12 years agoFix gdk_cairo_region_create_from_surface on big endian
Michel Dänzer [Sat, 19 Nov 2011 17:23:27 +0000 (12:23 -0500)]
Fix gdk_cairo_region_create_from_surface on big endian

gdk_cairo_region_create_from_surface doesn't work correctly on PPC.
This is most prominently seen with the GTK window resize grip, the
shape of which is mirrored every eight pixels horizontally.

At the same time, use an A1 surface for the resize grip shape to
eliminates an A8->A1 surface conversion.

12 years agowin32: Default to gtk-win32 theme
Alexander Larsson [Fri, 18 Nov 2011 15:53:18 +0000 (16:53 +0100)]
win32: Default to gtk-win32 theme

12 years agoBuild in default win32 css as gtk-win32
Alexander Larsson [Fri, 18 Nov 2011 15:52:38 +0000 (16:52 +0100)]
Build in default win32 css as gtk-win32

12 years agoBuild gtkwin32css.h from gtk-win32.css
Alexander Larsson [Fri, 18 Nov 2011 15:43:45 +0000 (16:43 +0100)]
Build gtkwin32css.h from gtk-win32.css

12 years agowin32-theme: Add gtk-win32.css
Alexander Larsson [Fri, 18 Nov 2011 15:20:32 +0000 (16:20 +0100)]
win32-theme: Add gtk-win32.css

12 years agogdk: remove the GET_EFFECTIVE_KEYMAP() stuff from gdkkeys-x11.c
Michael Natterer [Fri, 18 Nov 2011 15:42:25 +0000 (16:42 +0100)]
gdk: remove the GET_EFFECTIVE_KEYMAP() stuff from gdkkeys-x11.c

We don't support passing a NULL keymap any longer. Also pull some
precondition checks into the parent class.

12 years agogdk_x11_keymap_get_modifier_mask(): always return a value
Michael Natterer [Fri, 18 Nov 2011 14:26:09 +0000 (15:26 +0100)]
gdk_x11_keymap_get_modifier_mask(): always return a value

12 years agoTurn the private #define for the group-shifting modifier into API
Michael Natterer [Fri, 18 Nov 2011 14:14:31 +0000 (15:14 +0100)]
Turn the private #define for the group-shifting modifier into API

Add GDK_MODIFIER_INTENT_SHIFT_GROUP to enum GdkModifierIntent
and handle it in gdk_keymap_get_modifier_mask(). Add an X11
impl of the method and return keymap_x11->group_switch_mask.
Return 0 from the default impl because we don't know.

12 years agoBug 663856 - Make option-foo accelerators use the right symbol
Michael Natterer [Fri, 18 Nov 2011 11:25:03 +0000 (12:25 +0100)]
Bug 663856 - Make option-foo accelerators use the right symbol

If the keyboard group shifting modifier is *also* a normal
accelerator modifier, we need to special case it when calling
gdk_keymap_translate_keyboard_state(), so we get the right
key symbol for accelerators (for example we want Option-O,
not Option-Ø displayed in menu items). This patch should only
affect quartz where the Alt key both shifts the group and can
be used as accel modifier, and not X11 or Win32 where AltGr
is not used for accelerators.

- fix quartz' gdk_keymap_translate_keyboard_state() to return
  the right consumed_modifiers
- add _gtk_translate_keyboard_accel_state() which does the
  special casing
- use it everywhere instead of gdk_keymap_translate_keyboard_state()

12 years agoFix tiny docs typo.
Murray Cumming [Fri, 18 Nov 2011 10:30:07 +0000 (11:30 +0100)]
Fix tiny docs typo.

12 years agowin32-theme: Support mixing parts
Alexander Larsson [Fri, 18 Nov 2011 10:06:32 +0000 (11:06 +0100)]
win32-theme: Support mixing parts

We need this to do something about the non-existing inconsistent
radio button state in win32.

12 years agoAdd support for win32 theme colors as symbolic colors
Alexander Larsson [Fri, 18 Nov 2011 09:16:38 +0000 (10:16 +0100)]
Add support for win32 theme colors as symbolic colors

12 years agoAdd pulse style class and use in progress bar and entry
Alexander Larsson [Fri, 18 Nov 2011 09:15:04 +0000 (10:15 +0100)]
Add pulse style class and use in progress bar and entry

This is used for indeterminate progress reporting.

12 years agoSet top/bottom style classes on spinbutton buttons
Alexander Larsson [Thu, 17 Nov 2011 18:43:13 +0000 (19:43 +0100)]
Set top/bottom style classes on spinbutton buttons

12 years agoAdd top/left/bottom/right style classes to steppers
Alexander Larsson [Thu, 17 Nov 2011 16:34:05 +0000 (17:34 +0100)]
Add top/left/bottom/right style classes to steppers

This is needed for e.g. win32 theming, but is also generally
useful.

12 years agowin32-theme: Support multiple parts being combined
Alexander Larsson [Thu, 17 Nov 2011 13:34:39 +0000 (14:34 +0100)]
win32-theme: Support multiple parts being combined

This is needed for e.g. scrollbar sliders

12 years agoRender background image if set for checks and options
Alexander Larsson [Wed, 16 Nov 2011 16:03:09 +0000 (17:03 +0100)]
Render background image if set for checks and options

The default theme engine draws a fallback check/radio image, but
doesn't let you replace this. We now check if a background image
is set and if so render that instead of the default fallbacks.

12 years agowin32-theme: Support -gtk-win32-size CSS value
Alexander Larsson [Wed, 16 Nov 2011 11:19:20 +0000 (12:19 +0100)]
win32-theme: Support -gtk-win32-size CSS value

12 years agoAdd initial cut at win32 theme support for CSS
Alexander Larsson [Tue, 15 Nov 2011 19:52:45 +0000 (20:52 +0100)]
Add initial cut at win32 theme support for CSS

We now support -gtk-win32-theme-part(class,part,state) in background
and border-image CSS properties. This renders the corresponding
theme part using DrawThemeBackground() and acts as a base for a
CSS based windows theme.

Note that we build the parsing code even on non-win32 so that
all themese will parse the same on all arches. We draw pink instead
of the actual theme parts on non-win32 though.

12 years agomake _gtk_style_property_resolve copy the result to an output value
Alexander Larsson [Tue, 15 Nov 2011 16:15:49 +0000 (17:15 +0100)]
make _gtk_style_property_resolve copy the result to an output value

This is preparation for allowing it to return a newly created
value, rather than just copying one.

12 years agoActually pass in a StylePropertyContext when getting images
Alexander Larsson [Tue, 15 Nov 2011 16:13:22 +0000 (17:13 +0100)]
Actually pass in a StylePropertyContext when getting images

This is in preparation for getting size-dependent results back
with the win32 themes.

12 years agoRemove _gtk_style_properties_peek_property
Alexander Larsson [Tue, 15 Nov 2011 16:10:22 +0000 (17:10 +0100)]
Remove _gtk_style_properties_peek_property

We want to move to a world where getting style properties can
allocate a new value, depending on the style property context.
In this world we can't expose "peeking" a property, as we neet
to return a ref to the property that is newly created.

So, we move the peek code into get_property and use get_property
from get_valist.

12 years agoAdd GtkStylePropertyContext and use it in when getting style properties
Alexander Larsson [Tue, 15 Nov 2011 15:56:59 +0000 (16:56 +0100)]
Add GtkStylePropertyContext and use it in when getting style properties

At the toplevel we have _gtk_theming_engine_get, which lets us pass
in a property context with (atm) the size to get the property for.
Then there is a lot of plumbing to push this down into the lower
layers of the style property code until finally hitting
the property resolvers.

I need this because I will be adding a property resolver for win32
theme parts, and they render differently depending on the size
(i.e. they don't scale linearly). The idea is that the code
to get the background properties will pass in the final size
and we will resolve the theme part specification to that particular
size.

If the old non-context calls are used we just hardcode a size
of 100x100.

12 years agoSave a generic boxes source in GtkImageBorder
Alexander Larsson [Tue, 15 Nov 2011 15:26:46 +0000 (16:26 +0100)]
Save a generic boxes source in GtkImageBorder

We used to special handle gradient, but we want to be able to store
other sources (that will eventually resolve to cairo patterns).
For instance, this is needed to handle win32 theme part sources.

12 years agowin32: make fixup_event correctly ref all GdkWindow members
Alexander Larsson [Thu, 17 Nov 2011 11:25:56 +0000 (12:25 +0100)]
win32: make fixup_event correctly ref all GdkWindow members

This fixes some crashes related to owner changes and selections.

12 years agoref, don't unref, requestor member when copying GdkEvents
Alexander Larsson [Thu, 17 Nov 2011 10:59:27 +0000 (11:59 +0100)]
ref, don't unref, requestor member when copying GdkEvents

This seems like a typo, we should obviously ref the member when
copying. Apparently nobody copied GDK_SELECTION_* events...

12 years agonotebook: don't leak the action widgets
Michael Natterer [Wed, 16 Nov 2011 23:49:42 +0000 (00:49 +0100)]
notebook: don't leak the action widgets

The notebook owns them, so it must destroy them in destroy().
(cherry picked from commit 87ebfb19fd9422de1d778396770c02e91e6481e8)

12 years agoviewport: Get rid of deprecation warnings
Benjamin Otte [Wed, 16 Nov 2011 23:29:30 +0000 (00:29 +0100)]
viewport: Get rid of deprecation warnings

12 years agoviewport: Don't use deprecated functions
Benjamin Otte [Wed, 16 Nov 2011 23:27:35 +0000 (00:27 +0100)]
viewport: Don't use deprecated functions

12 years agoviewport: adjustments are never NULL
Benjamin Otte [Wed, 16 Nov 2011 23:25:15 +0000 (00:25 +0100)]
viewport: adjustments are never NULL

So there's no need to check for that in the getters.

12 years agoviewport: Initialize the adjustment in the init function
Benjamin Otte [Wed, 16 Nov 2011 23:24:15 +0000 (00:24 +0100)]
viewport: Initialize the adjustment in the init function

12 years agoviewport: Don't unset the adjustment on finalize
Benjamin Otte [Wed, 16 Nov 2011 23:20:09 +0000 (00:20 +0100)]
viewport: Don't unset the adjustment on finalize

We unset it on destroy already. This makes it consistent with all other
scrollables.

12 years agotreeview: fix the build
Xan Lopez [Wed, 16 Nov 2011 21:41:57 +0000 (22:41 +0100)]
treeview: fix the build

12 years agotreeview: use gtk_tree_view_ensure_background() consistently
Cosimo Cecchi [Wed, 16 Nov 2011 21:01:36 +0000 (16:01 -0500)]
treeview: use gtk_tree_view_ensure_background() consistently

In gtk_tree_view_state_flags_changed() we were setting the background on
the bin_window without the necessary "view" style class, making the
treeview render with the wrong color in some circumstances (such as when
adding an empty treeview in a window, as spotted by Brian Smith).

12 years agotreeview: factor out gtk_tree_view_ensure_background()
Cosimo Cecchi [Wed, 16 Nov 2011 20:59:47 +0000 (15:59 -0500)]
treeview: factor out gtk_tree_view_ensure_background()

We'll use it shortly.

12 years agoMaintain map/unmap invariants in GtkRecentChooserDialog
Kjell Ahlstedt [Wed, 16 Nov 2011 08:03:12 +0000 (09:03 +0100)]
Maintain map/unmap invariants in GtkRecentChooserDialog

We used to explicitly map and unmap the child GtkRecentChooserWidget when
mapping and unmapping the dialog, respectively. Now that GtkContainer actually
unmaps child widgets (instead of avoiding that), we can assume that the
child GtkRecentChooserWidget will be unmapped when we want it to be.

This fixes a warning from gtk_widget_verify_invariants(), as we were mapping
our child widget before calling our parent class' ::map() handler. Bug #659257.

12 years agotests: Add tests for keypad keys parsing
Bastien Nocera [Wed, 16 Nov 2011 18:49:54 +0000 (18:49 +0000)]
tests: Add tests for keypad keys parsing

See https://bugzilla.gnome.org/show_bug.cgi?id=606727

12 years agogtk: Parse keypad numbers correctly
Bastien Nocera [Wed, 16 Nov 2011 18:47:39 +0000 (18:47 +0000)]
gtk: Parse keypad numbers correctly

Try to find keycodes for keysyms at level-0 in the keymap, but don't
fail to parse if we can find those in level-1. This fixes the
parsing of numerical keys on the keypad, which require a shift level
through Num-Lock.

https://bugzilla.gnome.org/show_bug.cgi?id=606727

12 years agotests: Check "without keycode" code path too
Bastien Nocera [Wed, 16 Nov 2011 16:13:01 +0000 (16:13 +0000)]
tests: Check "without keycode" code path too

And see whether the output of gtk_accelerator_parse()
matches that of gtk_accelerator_parse_with_keycode()

12 years agoRevert "tests: Disable accel test until fixed"
Bastien Nocera [Wed, 16 Nov 2011 15:02:06 +0000 (15:02 +0000)]
Revert "tests: Disable accel test until fixed"

The test now works correctly.

This reverts commit ba397a3431ee3850a72d8eab394b7ee62e65b737.

12 years agoUpdate comment to new realities
Matthias Clasen [Wed, 16 Nov 2011 05:27:14 +0000 (00:27 -0500)]
Update comment to new realities

12 years agoa11y: Merge function into only caller
Benjamin Otte [Wed, 16 Nov 2011 02:52:02 +0000 (03:52 +0100)]
a11y: Merge function into only caller

Also cleann up comments that are misleading now.

12 years agotreeview: Call a11y functions for culmn changes directly
Benjamin Otte [Wed, 16 Nov 2011 02:47:33 +0000 (03:47 +0100)]
treeview: Call a11y functions for culmn changes directly

This way, the a11y code knows if a column was reordered, added or
removed and can do the right things instead of trying to guess and
getting it wrong.

Also, this patch finalizes the changes so that only visible columns
exist to the accessibility interface.

12 years agoa11y: Unify column-by-index getting
Benjamin Otte [Wed, 16 Nov 2011 01:40:49 +0000 (02:40 +0100)]
a11y: Unify column-by-index getting

We are only ever interested in visible columns. Invisible columns do not
exist for all a11y cares.

12 years agotests: Add an invisible column to the tree test
Benjamin Otte [Wed, 16 Nov 2011 00:32:07 +0000 (01:32 +0100)]
tests: Add an invisible column to the tree test

This one should be invisible if everything works as it should.

12 years agotests: Add checks for row/column index to accessibility-dump
Benjamin Otte [Tue, 15 Nov 2011 21:29:45 +0000 (22:29 +0100)]
tests: Add checks for row/column index to accessibility-dump

12 years agotests: Don't dump stuff twice for tables
Benjamin Otte [Tue, 15 Nov 2011 21:18:40 +0000 (22:18 +0100)]
tests: Don't dump stuff twice for tables

Captions and headers are printed as part of the children, so we can
avoid the duplication by just printing the name.

12 years agoa11y: Use gtk_tree_view_get_n_columns()
Benjamin Otte [Tue, 15 Nov 2011 15:45:48 +0000 (16:45 +0100)]
a11y: Use gtk_tree_view_get_n_columns()

In fact, invent our own function get_n_columns() which actually only
counts the visible functions, because that's the only ones we care
about.

The places where it's not used yet will be changed in the coming
patches.

12 years agoAPI: treeview: Add gtk_tree_view_get_n_columns()
Benjamin Otte [Tue, 15 Nov 2011 15:37:24 +0000 (16:37 +0100)]
API: treeview: Add gtk_tree_view_get_n_columns()

12 years agoa11y: Fix get_column_index()
Benjamin Otte [Tue, 15 Nov 2011 20:30:46 +0000 (21:30 +0100)]
a11y: Fix get_column_index()

12 years agoa11y: remove argument from get_column_number()
Benjamin Otte [Tue, 15 Nov 2011 19:24:51 +0000 (20:24 +0100)]
a11y: remove argument from get_column_number()

The argument was always FALSE.

12 years agotreeview: Simplify a check
Benjamin Otte [Tue, 15 Nov 2011 14:12:34 +0000 (15:12 +0100)]
treeview: Simplify a check

Use the same method as elsewhere.

12 years agotreeview: Minor beautification
Benjamin Otte [Tue, 15 Nov 2011 14:04:20 +0000 (15:04 +0100)]
treeview: Minor beautification

12 years agotests: Up the ante
Benjamin Otte [Sun, 13 Nov 2011 03:43:32 +0000 (04:43 +0100)]
tests: Up the ante

Put 10000 rows into the tree-performance test, so the tests don't finish
in <1s. That's too little for profiling.

12 years agoa11y: remove useless argument
Benjamin Otte [Sun, 13 Nov 2011 03:28:52 +0000 (04:28 +0100)]
a11y: remove useless argument

The argument is unused these days, so don't confuse people with it.

12 years agoa11y: Remove clean_rows function
Benjamin Otte [Sun, 13 Nov 2011 01:52:04 +0000 (02:52 +0100)]
a11y: Remove clean_rows function

It doesn't do anything anymore.

12 years agoa11y: Don't crete cell infos for non-children
Benjamin Otte [Sun, 13 Nov 2011 01:47:42 +0000 (02:47 +0100)]
a11y: Don't crete cell infos for non-children

When we use a container inbetween, this ensures that the cell infos
don't replace each other in the hash table.

12 years agoa11y: Make find by index lookup use the hash table again
Benjamin Otte [Sat, 12 Nov 2011 05:37:08 +0000 (06:37 +0100)]
a11y: Make find by index lookup use the hash table again

This should give back the performance that was lost with the commit
"a11y: Make the cache lookup function inefficient".

12 years agoa11y: Add a hash function for cell infos
Benjamin Otte [Sat, 12 Nov 2011 04:38:52 +0000 (05:38 +0100)]
a11y: Add a hash function for cell infos

Note that comparing the tree is not necessary as the nodes are already
unique per row.

12 years agoa11y: Compute the cell index quicker
Benjamin Otte [Sat, 12 Nov 2011 04:24:32 +0000 (05:24 +0100)]
a11y: Compute the cell index quicker

We have the node available, so we can compute it from there.

12 years agotreeview: Add _gtk_rbtree_node_get_index()
Benjamin Otte [Sat, 12 Nov 2011 04:15:53 +0000 (05:15 +0100)]
treeview: Add _gtk_rbtree_node_get_index()

.. as a replacement for _gtk_rbtree_node_find_parity(). Instead of 1 or
0, the function now returns the index of node in the complete tree
(counting from the root). And this is of course identical to the row
number.

12 years agoa11y: Compute index the easy way
Benjamin Otte [Sat, 12 Nov 2011 04:05:19 +0000 (05:05 +0100)]
a11y: Compute index the easy way

index is row_index * n_columns + column_index

We use it everywhere, why not here?

12 years agoa11y: Redo indexing
Benjamin Otte [Sat, 12 Nov 2011 03:56:29 +0000 (04:56 +0100)]
a11y: Redo indexing

We now index rows by actual expanded row number and don't count them
ourselves no more.

12 years agotests: Ensure the treeview is expanded
Benjamin Otte [Sat, 12 Nov 2011 03:55:32 +0000 (04:55 +0100)]
tests: Ensure the treeview is expanded

Otherwise the test code might ignore collapsed rows, which makes it go
rather fast.

12 years agotreeview: Redo row tracking
Benjamin Otte [Sat, 12 Nov 2011 02:39:48 +0000 (03:39 +0100)]
treeview: Redo row tracking

Track the RBNode/RBTree instead of keeping a TreeRowReference. This is a
whole lot faster and less error-prone.

Also, notify the accessible of removal of rows before actually removing
them, so we have a chance to clean up.

12 years agowidget: Add _gtk_widget_peek_accessible()
Benjamin Otte [Sat, 12 Nov 2011 02:38:40 +0000 (03:38 +0100)]
widget: Add _gtk_widget_peek_accessible()

This function returns the accessible if it already exists. This way we
can call functions on the accessible from the widget itself instead of
having to rely on signals.

12 years agoa11y: Remove needless checks
Benjamin Otte [Sat, 12 Nov 2011 02:20:12 +0000 (03:20 +0100)]
a11y: Remove needless checks

The cell_row_ref and cell_col_ref variables are always non-NULL, so
there's no need to check it.

12 years agoa11y: Add a cell_info_get_path() function
Benjamin Otte [Sat, 12 Nov 2011 02:18:13 +0000 (03:18 +0100)]
a11y: Add a cell_info_get_path() function

Reads better and we only need to change it in one place.

12 years agotreeview: Remove unnecessary variable
Benjamin Otte [Sat, 12 Nov 2011 01:18:26 +0000 (02:18 +0100)]
treeview: Remove unnecessary variable

tmptree has the same value as tree everywhere, so just use tree instead.

12 years agoa11y: Mark object as defunct in destroy notify
Benjamin Otte [Fri, 11 Nov 2011 20:03:49 +0000 (21:03 +0100)]
a11y: Mark object as defunct in destroy notify

Instead of requiring a special functio to be called before just removing
from the hash table. This simplifies code a lot that used to remove from
the hash table.

12 years agoa11y: Remove in_use member
Benjamin Otte [Fri, 11 Nov 2011 19:30:39 +0000 (20:30 +0100)]
a11y: Remove in_use member

After the latest changes, it is now always TRUE, so adjust the code
accordingly.

12 years agoa11y: Don't garbage collect anymore
Benjamin Otte [Fri, 11 Nov 2011 19:25:14 +0000 (20:25 +0100)]
a11y: Don't garbage collect anymore

Instead, remove cell infos immediately from the hash table.

12 years agoa11y: Improve cell_infos table
Benjamin Otte [Fri, 11 Nov 2011 19:18:09 +0000 (20:18 +0100)]
a11y: Improve cell_infos table

- Name it properly
- Make it just a set of cell_infos

Currently it's using direct hash, but as long as we don't actually
lookup stuff from it, who cares...

12 years agoa11y: Implement find_cell_info() using qdata lookup
Benjamin Otte [Fri, 11 Nov 2011 19:11:50 +0000 (20:11 +0100)]
a11y: Implement find_cell_info() using qdata lookup

That should be orders of magnitude faster then iterating over a hash
table.

12 years agoa11y: Store the cellinfo in the cell's qdata
Benjamin Otte [Fri, 11 Nov 2011 19:06:20 +0000 (20:06 +0100)]
a11y: Store the cellinfo in the cell's qdata

12 years agoa11y: Remove index from cellaccessible
Benjamin Otte [Fri, 11 Nov 2011 02:02:31 +0000 (03:02 +0100)]
a11y: Remove index from cellaccessible

That way we also get rid of the refresh_index function.

12 years agoa11y: Don't use the cell's index anymore
Benjamin Otte [Fri, 11 Nov 2011 01:52:47 +0000 (02:52 +0100)]
a11y: Don't use the cell's index anymore

We implement get_cell_index() now, so it's no longer necessary.
As a side effect, we need a different index for our hash table (which
now has a wrong name, but that will soon change).

12 years agoa11y: Implement get_child_index in treeview
Benjamin Otte [Fri, 11 Nov 2011 01:48:40 +0000 (02:48 +0100)]
a11y: Implement get_child_index in treeview

12 years agoa11y: Implement get_child_index in treeviewaccessible
Benjamin Otte [Fri, 11 Nov 2011 01:19:28 +0000 (02:19 +0100)]
a11y: Implement get_child_index in treeviewaccessible

12 years agogtk: Fix compiler warnings from include fixes
Benjamin Otte [Fri, 11 Nov 2011 00:43:54 +0000 (01:43 +0100)]
gtk: Fix compiler warnings from include fixes