In pygtk custom cellrenderers were done inheriting
gtk.GenericCellRenderer, and overriding the on_* methods. Now we
inherit Gtk.CellRenderer and the methods to override changed to do_* .
The destroy signal was moved to Gtk.Widget [1] so the Gtk.CellRenderer
doesn't have it anymore. Now we do the cleanup in the python
destructor method. A testcase tests/graphics/customdestroy.py shows
how is done.
tests/graphics/cellrenderericon.py tests the usage.
Signed-off-by: Manuel Quiñones <manuq@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
[1] http://developer.gnome.org/gtk3/stable/ch24s02.html#id1459754
Extending the generic list will make those mime-types available in the
filter list (ObjectChooser, Journal). We have been extending the generic
list similarly before [1][2].
For the longer term fix see [3] to add support for custom mime types in
the ObjectChooser.
[1] http://bugs.sugarlabs.org/ticket/1319
[2] http://bugs.sugarlabs.org/ticket/1340
[3] http://bugs.sugarlabs.org/ticket/834
Signed-off-by: Manuel Kaufmann <humitos@gmail.com>
Acked-by: Simon Schampijer <simon@laptop.org>
Draw a black background in the buttons when the palette is up, as
commit 01a06943 did with the ToolButton. As the comment for that
commit states, we can change the prelight background color in the
theme, but not when the mouse moves over the Palette.
Also add testcase to test the three toolbuttons.
Signed-off-by: Manuel Quiñones <manuq@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
When git is installed, it will return non-zero if it gets asked to list
the files in a non-git-repository. The subprocess.Popen instantiation is
successful in this case and the returncode attribute will contain the
error code from git. The current code handles this fine and does fall
back to our own source file listing facility.
If git isn't installed, however, trying to instantiate subprocess.Popen
will fail with OSError. We need to catch this and fall back to our own
source file listing facility like we do for the non-repository case.
Signed-off-by: Simon Schampijer <simon at laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Manuel Quiñones <manuq@laptop.org>
To the grab_keys function we pass a list of strings the key
grabber should listen for. In order to make this
introspectable we need to annotate the function correctly. I
followed the instructions for passing arrays [1] and have
chosen the type of the array data looking at [2].
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Daniel Narvaez <dwnarvaez@gmail.com>
Tested-by: Manuel Quiñones <manuq@laptop.org>
[1] https://live.gnome.org/PyGObject/IntrospectionPorting#Passing_arrays
[2] https://live.gnome.org/GObjectIntrospection/Annotations#Default_Basic_Types
Use clicked instead of button-pressed-event, because button-pressed-event
and the button-release-event are not emitted on a touchscreen device by
a touch in GTK+ 3.
Signed-off-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Manuel Quiñones <manuq@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
This patch solves the following problems:
* API changed in the drag and drop code in Gtk.
Drag and drop is not working yet (SL #3796)
but the code needed to enable drag and drop is ported.
* Changes in the way to get color information from the theme
this is because Gtk.Style was deprecated by Gtk.StyleContext.
* The internal button was not visible.
Signed-of-by: Gonzalo Odiard <gonzalo@laptop.org>
-----
v2: Fixed comment based on manuq feedback, and add a note
about the non working drag and drop.
The icon consists of an GtkEventBox and an IconBuffer. The
GtkEventBox is a subclass of GtkBin which has its own window and
therefor is used to catch events for our IconBuffer which does
not have it's own window. The window of the EventBox is
made invisible.
The EventIcon does not emit the 'activated' signal when clicked
anymore, you can listen to 'button-release-event' to know
when the icon has been clicked. The EventIcon uses the
CursorInvoker to invoke a palette the same way as the
CanvasIcon did.
We keep the same API as with the CanvasIcon, only the 'size'
property is changed to be called 'pixel_size' in order to
make clearer which values it expects to be passed. We don't
expect a GtkIconSize to be passed here.
Another option would have been to put a SugarIcon inside a
a GtkEventBox and make the properties available through an
icon property but the API would have not been as nice and
logically it seems to make more sense to have the IconBuffer
being the base for both the SugarIcon and the SugarEventIcon.
This patch has ben developed based on the one that is
used in the shell port.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewd-by: Benjamin Berg <benzea@sugarlabs.org>
This code was still using regular Telepathy properties to
set important configuration such as Anonymous=False.
However, as of Telepathy specification 0.24.0, these properties have
gone away.
http://telepathy.freedesktop.org/spec/Channel_Type_Text.html
Changed in 0.24.0. This interface used to have a bunch of clunky
Telepathy.Properties. They have been removed in favour of D-Bus
properties on the Room2, Subject2 and RoomConfig1 interfaces.
Switch to using RoomConfig1 (where available) to set this
configuration. The invite-restricted flag is no longer available and
actually seems to have been removed a long while back.
Fixes sharing of activities over gabble on new platforms such
as Fedora 17.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Sascha Silbe <silbe@activitycentral.com>
The create_palette mechanism allows to create
palettes on demand and not for each icon upfront
whether it will be needed or not.
If you hover over an EventIcon and there is no
palette already associated with the Invoker, the Invoker
will call create_palette to see if it can be created
on demand. With this patch the EventIcon will return None
here (see as well CellRendererIcon or ToggleToolButton which
are having the same default behavior).
When subclassing EventIcon the create_palette method can
be overwritten and a Palette returned (see for example
the ActivityIcon in the HomeView).
Without this patch you can see tracebacks when hovering
over the EventIcon because the Invoker tries to call
create_palette.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Benjamin Berg <benzea@sugarlabs.org>
Acked-by: Manuel Quiñones <manuq@laptop.org>
Position the event box window above the windows of its child, that way
all events inside the event box will go to the event box. If the window is
below, events in windows of child widgets will first go to that widget,
and then to its parents [1].
[1] http://developer.gnome.org/gtk3/3.4/GtkEventBox.html#gtk-event-box-set-above-child
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Benjamin Berg <benzea@sugarlabs.org>
Acked-by: Manuel Quiñones <manuq@laptop.org>
We do use the EventBox only to receive events, hence the
window that the even box creates should be a GDK_INPUT_ONLY
window, which means that it is invisible and only serves to
receive events [1].
[1] http://developer.gnome.org/gtk3/3.4/GtkEventBox.html#gtk-event-box-set-visible-window
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Benjamin Berg <benzea@sugarlabs.org>
Acked-by: Manuel Quiñones <manuq@laptop.org>
The _ToolbarPalette does get passed the invoker on
initialisation. But we do create the PaletteWindowWidget
later. We do attach the invoker to the widget when calling
_setup_widget that is why it was still working without that
patch.
This patch prevents the traceback that we had because of not
having a widget at this point.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Tested-by: Manuel Quiñones <manuq@laptop.org>
set_data/get_data not available anymore is not available anymore [1]. The
recommended approach is using a python attribute, which we do.
Changing to use the attribute in the activity class slipped by mistake
into 6330204e91.
[1] https://bugzilla.gnome.org/show_bug.cgi?id=641944
Signed-off-by: Simon Schampijer <simon@laptop.org>
Tested-by: Manuel Quiñones <manuq@laptop.org>
Must be done early, some activities set translations globally. Remove
the support for the langpackdir.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Sugar ToolButton inherits from Gtk.ToolButton but is not overwriting
'icon_name' property. So 'icon_name' can be passed to the constructor
of Sugar ToolButton but the result is different than setting it via
ToolButton.props.icon_name.
Signed-off-by: Manuel Quiñones <manuq@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
In GDK3 the 'xid' attribute is gone and is replaced with
gdk_x11_window_get_xid(), or the get_xid() method in Python.
Needed to be able to open the object chooser from Browse without hassle.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
The minimum height/width requested by the GtkViewport still tries to cater
for all contained children, which makes the [VH]Tray widgets to grow as
new items are added. Instead, request a minimum width/height of 0 to avoid
the Tray from growing, and having scrolling kick in instead.
Also, fixed what seemed to be a typo in do_get_preferred_height(), where
the viewport width was requested instead.
http://bugs.sugarlabs.org/ticket/3522
Signed-off-by: Carlos Garnacho <carlos@lanedo.com>
Acked-by: Simon Schampijer <simon@laptop.org>
The accelerator in the primary information in the Palette
has not been drawn because there was not enough space
reserved for it. The preferred size we get back for the
Palette window does not include the accelerator of the
Gtk.AccelLabel. We need to include that in our calculation for
the Palette size.
In order to make that information available which is part
of the Palette class we need to pass the instance to the
PaletteWindowWidget instance.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
We use json as included in Python. This will make Sugar dependent on
Python 2.6 and Python 2.7 to have the highest JSON performance.
Signed-off-by: Simon Schampijer <simon@laptop.org>
gobject introspection bindings for librsvg have been pushed to librsvg
master [1] in 2.35.0, which solved [2]. We only have slight adopts to
make in our usage, for example we can not pass the data property
to the default constructor anymore and get_width and get_height is not
available anymore for the handle, but we can use the properties
instead.
The sugar-toolkit-gtk3 and therefore Activities that have been ported
to it do need a version of librsvg >= 2.35.0 to be able to work,
which ships for example with Fedora 17.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
[1] http://git.gnome.org/browse/librsvg/
[2] https://bugzilla.gnome.org/show_bug.cgi?id=663049
[3] http://developer.gnome.org/rsvg/stable/RsvgHandle.html
First we needed to port the Palette code to use a minimum size. The default size
is two times the GRID_CELL_SIZE. Since the request-phase of the traditional GTK+
geometry management has been replaced by a height-for-width system [1] we have
to compensate for that. Furthermore we need to pass the invoker from the
PaletteWindow to the _PaletteWindowWidget for the gap calculation code for
drawing the border around the Palette.
We do the drawing of the border for the toolbutton in the base class, moved
this from the ToolbarButton and made sure we are drawing in the right order.
In the ToolButton we draw as well a black background for the ToolButton when
the Palette is up. While the mouse is over the button we can do that in the
theme, but not when the mouse moves over the Palette.
[1] http://developer.gnome.org/gtk3/3.0/ch25s02.html#id1525688
Signed-off-by: Simon Schampijer <simon@laptop.org>
This draws the grey line around the toolbutton icon with a gap at the
bottom and a grey line at the top of the subtoolbar. Furthermore it
gets the highlightning of the button correct, in the pressed and hover
state. This patch depends on the sugar-artwork patch with the id
7464b808eb12b1df650952e3c8214acff1d1360f.
Signed-off-by: Gonzalo Odiard <gonzalo@laptop.org>
Signed-off-by: Manuel Quiñones <manuq@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
Previously we were using a GdkPixmap [1] created with the get_snapshot [2]
method of the Gtk.Widget. GTK 3 encourages to use cairo surfaces
now instead.
The ported mothod does create a cairo surface similar to that of the canvas'
window and draws on that. Then we create a cairo image surface with
the desired preview size and scale the canvas surface on that.
Several people have been involved in this work: Gonzalo Odiard,
Manuel Quiñones and Benjamin Berg.
[1] http://developer.gnome.org/gdk/stable/gdk-Bitmaps-and-Pixmaps.html#GdkPixmap
[2] http://developer.gnome.org/gtk/stable/GtkWidget.html#gtk-widget-get-snapshot
With 820efa56b9
gtk.gdk.x11_get_server_time(window) wasn't correctly converted
to GdkX11.x11_get_server_time(window). Found when tesing
collaboration. Opened 669264 for the upstream fix of
pygi-convert.sh.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
With the new 'Write to Journal anytime' feature it is possible
to write a description within an activity. Hopefully this leads
to more usage of the Journal for reflection by the learner. The
hinting nature ('you have not named your session yet') of the
Naming Alert has not been replaced, though.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Reviewed-by: Manuel Quiñones <manuq@laptop.org>
This is the implementation of the 'Write to Journal anytime'
feature [1].
The patch itself adds a DescriptionItem to the activity
sub-toolbar to make editing a Journal entry description
from within the activity possible. The code has the same
error handling as the TitleEntry.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Signed-off-by: Walter Bender <walter.bender@gmail.com>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Reviewed-by: Manuel Quiñones <manuq@laptop.org>
[1] http://wiki.sugarlabs.org/go/Features/Write_to_journal_anytime
- removed deprecated imports from the activity module, use the
widgets module instead
- removed the ActivityToolbox class
- removed the Stop button from the ActivityToolbar
- removed set_toolbar/get_toolbar API from the window module
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
This API is not in use anywhere in the shell and has
been deprecated for quite some time.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
The objectchooser had the 'title', 'flags' and 'buttons' parameters
deprecated for a long time, remove them now completely. The
only parameters allowed are now the 'parent' and the 'what_filter'.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
Use the get_shared_activity method instead
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Daniel Drake <dsd@laptop.org>
see 0082e10f8e for
the complete reasoning.
e022aa8e4a already made the button
invisible, keeping it only for backwards compatibility.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-By: Sascha Silbe <silbe@activitycentral.com>
Acked-by: Daniel Drake <dsd@laptop.org>
We need set the theme in gsettings according the scale in
the SUGAR_SCLING environment variable.
Signed-off-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
Moving from GTK2 to GTK3 has presented various challenges regarding
palettes.
In GTK2, we were able to access some internal API of the GtkMenu class
and use it to embed a GtkMenu in a regular window. As of GTK3, that API
has become private and we can no longer access it.
We still want to use GtkMenu for the advanced functionality it provides
(multiple-level menus, keyboard navigation, etc), but we are now limited
to popping it up with its own (internal) window, rather than being able
to pack it into one of our own.
Our palettes can historically be used either as a menu, or as a general
area where widgets can be added, or both. The new restrictions upon
GtkMenu force some changes here, but we work hard to stick to the old
API as far as possible.
A Palette instance now acts as a controller of either a "window widget"
(where any type of widget can be displayed as usual) or a "menu widget"
which just pops up a GtkMenu. A Palette defaults to the window mode, but
dynamically switches to menu mode if/when the user attempts to access
the menu element.
As a result of this, palettes can now pack either a user-defined collection
of widgets, or a menu, but types can no longer be mixed. This should
only affect a handful of palettes which will need to pick a single
approach and convert to it.
Some further challenges are presented by the fact that GtkMenu performs a
grab on the whole screen, meaning that all input events are delivered to
the GtkMenu widget. Through some careful event filtering and examination
of the mouse cursor position we are still able to determine when the mouse
has entered or left the invoker or menu areas.
This work is authored by Benjamin Berg, Marco Pesenti Gritti, Simon
Schampijer and Daniel Drake.
Fix some trivial issues missed earlier: various missing imports,
some minor API changes to adapt to, do_size_request simple porting,
etc.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CanvasIcon and CanvasInvoker were removed in a previous GTK3-porting commit
as they were based on hippocanvas.
However, this leaves the toolkit with some missing functionality:
there is no longer a trivial way to show an icon which can receive mouse
events and pop up a palette. Such functionality is used in various
places throughout the shell and activities.
Reimplement this functionality as EventIcon and CursorInvoker.
Instead of reimplementing much of the Icon class (like CanvasIcon did),
EventIcon opts for a more simplistic encapsulation of an Icon object.
This means trivial API changes for CanvasIcon users who must now
use the 'icon' property with the Icon API.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
Now that we avoid linking with pygtk2/pygobject2, we need to remove
this initialisation call so that the module can be loaded at runtime.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
In GTK2, theme name and icon theme name properties could be set in the
GTK2 RC file, at runtime, or by the X settings daemon.
For GTK3, the RC file configuration route for these settings has been
removed. As we do not currently have a settings daemon implementation,
apply these important settings at runtime, early in the Activity
class.
Signed-off-by: Daniel Drake <dsd@laptop.org>
As we move to adding support for a second UI toolkit (GTK+ 3.x),
the sugar-activity binary used by all activities must become
backend-toolkit-independent. It would be wasteful to have two backend
toolkits loaded in memory, and in the GTK2/GTK3 case, it is impossible
(importing both results in an instant crash).
To achieve this, we split the existing sugar-toolkit activity/main.py:main()
functionality into two parts, moving it into the sugar-activity binary and
the Activity class as follows:
1. All toolkit-specific stuff is moved into the Activity class (i.e.
everything that interacts with GTK)
2. Everything that can be reasonably/easily moved into the Activity class
is also moved.
3. What remains is the stuff that is inherently involved with the
construction of the Activity object, not related to UI toolkits. This
is moved into the sugar-activity binary.
main.py is then removed from sugar-toolkit, and sugar-activity is moved
from sugar to sugar-toolkit-gtk3 in order to keep toolkit-related code
with the toolkit itself.
With this work done, the one remaining question is how to invoke the main
loop. An optional run_main_loop() method is added to the activity class,
for GTK2 this will run the GTK2 main loop, for GTK3 the GTK3 main loop will
be run, etc.
Signed-off-by: Daniel Drake <dsd@laptop.org>
GTK3 removed some previously-deprecated API that we still use. This
includes GDK_DISPLAY(), gdk_x11_drawable_get_xdisplay(), and some
key constants.
Port our code to the new API.
[split patch into several parts, added minimal description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Probably needs cleaning up a bit. And we use pygtk-codegen, ugh...
This is the commit id when we imported sugar-base:
b9406e5c9c9df5404c5b0d995178b5edb4d93628
Signed-off-by: Daniel Drake <dsd@laptop.org>
[squashed two patches into one]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
In Pango 'Pango.attr_weight_new' is not yet introspectable [1].
[1] https://bugzilla.gnome.org/show_bug.cgi?id=646788
Signed-off-by: Simon Schampijer <simon@schampijer.de>
[changed description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
GtkWidget "expose-event" signal has been replaced by
a new "draw" signal [1]. The context is already
clipped [2], so do not base it on the values returned by
get_allocation like before.
[1] http://developer.gnome.org/gtk3/3.0/ch25s02.html#id1467092
[2] http://developer.gnome.org/gtk3/3.0/GtkWidget.html#GtkWidget-draw
Signed-off-by: Simon Schampijer <simon@schampijer.de>
[squashed with a patch by Benjamin Berg <benjamin@sipsolutions.net>;
removed useless additions]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
With PyGTK, all parameters of the Alignment constructor had defaults [1].
With GTK3+pygi, when using the explicit constructor (Alignment.new() resp.
gtk_alignment_new() [2]), all values would need to be passed. However when
using the GObject constructor, named properties can be passed in instead and
we only need to pass those that different from the default.
[1] http://developer.gnome.org/pygtk/stable/class-gtkalignment.html#constructor-gtkalignment
[2] http://developer.gnome.org/gtk/stable/GtkAlignment.html#gtk-alignment-new
Signed-off-by: Simon Schampijer <simon@schampijer.de>
[assembled from several patches; replaced description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
The following data fields that were provided by PyGTK are not accessible
directly in GTK3+pygi and need to be replaced by accessor calls:
Adjustment.lower
Adjustment.page_size
Adjustment.upper
Adjustment.value
Bin.child
Widget.parent
Widget.style
Widget.window
Based on patches by Daniel Drake <dsd@laptop.org>.
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
The previous set_source_pixbuf() invocation (on a Cairo context) involves
Gdk data types, so in the new introspection world we need to call a Gdk
function rather than operating on the Cairo object (even if Cairo had already
been converted to introspection).
Signed-off-by: Daniel Drake <dsd@laptop.org>
[added description; split out from another patch]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
GTK3 has replaced [1] the GTK2 geometry management with Height-for-width
Geometry Management [2]. This means we need to replace size_request() methods
with get_preferred_{width,height}().
[1] http://developer.gnome.org/gtk3/3.0/ch25s02.html#id1525688
[2] http://developer.gnome.org/gtk3/3.0/GtkWidget.html#geometry-management
Signed-off-by: Benjamin Berg <benjamin@sipsolutions.net>
[assembled from several patches; fixed up left-over plus sign; added
description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
This makes sugarext accessible through introspection. It is used
from the shell (key grabber, sound volume management) and the
shell session management.
Making the sugarext introspectable was done following the
descriptions at: http://live.gnome.org/GObjectIntrospection#Using_GI
Signed-off-by: Simon Schampijer <simon@schampijer.de>
This is only on a best-effort basis; the code will be in a broken state after
this patch and need to be fixed manually.
The purpose of committing the intermediate, non-working output is to make it
reproducible. It's impractical to manually review the changes.
The exact version used was 4f637212f13b197a95c824967a58496b9e3b877c from the
main pygobject repository [1] plus a custom patch [2] that hasn't been sent
upstream yet.
[1] git://git.gnome.org/pygobject
[2] https://sascha.silbe.org/patches/pygobject-convert-sugar-20111122.patch
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Some parts of GTK moved to GDK, so we need to import the latter for things to
work after the conversion script runs.
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
hippo-canvas isn't available in the GTK3 world, so we need to remove
anything that depends on it. Activities that still use it will need replace
hippo-canvas based widgets with native GTK ones before they can be ported to
GTK3.
[replaced description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
hippo-canvas isn't available in the GTK3 world and we can do fine without it
for rendering pixbufs.
[split out from another patch; added description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Make sure can_close_tabs can be passed in as a keyword parameter (to be set
by the GObject constructor).
[split out from another patch]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
8f1a821d68 only changed imports starting with
"import", but not those using "from ... import ...".
[replaced description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
The old gtk-2 based module will be present in
the 0.94 branch in the sugar-toolkit.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-by: Sascha Silbe <silbe@activitycentral.com>
Follow up patch for b582736375
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Daniel Drake <dsd@laptop.org>
Acked-by: Sascha Silbe <silbe@activitycentral.com>
Due to the removal of the manifest support we lost as well
the part of the code that does build the locale. This patch
adds that functionality back.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Remove an unnecessary toplevel widget.
This event box was originally added by Marco "to make it easier
to take screenshot of the canvas area only" (7f731457c2) but we're
unsure why this is, and it doesn't seem to be needed for our current
screenshot-taking mechanism. Screenshots continue to work fine after
removing this.
We need to set the title when we join a shared session AND when
we run a private session.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Tested-By: Gonzalo Odiard <gonzalo@laptop.org>
Reviewed-By: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Sascha Silbe <silbe@activitycentral.com>
There had been reports about Activities that had unexpected files in
the xo bundle (e.g. patches). There have been a recent change that the support
for the MANIFEST has been removed from the bundle builder. The MANIFEST
'controlled' which files were bundled. The code did include all the
files present in the folder (which includes patches etc).
The patch does use git-ls to get a list of files to include (like the
tarball packager). Furthermore the locale folder is included which has been
generated. Due to the API freeze we made '_get_files_in_git' a private method
which adds a bit of duplication to the 'get_files' method in the tarball
packager. A patch for master which will implement Builder.get_files_git(),
that can be used by XOPackager and SourcePackager.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-By: Daniel Drake <dsd@laptop.org>
In order to allow activities keeping their backward compatibility
we keep the Keep button around but do not add it to the toolbar. This
fixes activities that either remove or hide the Keep button or
alter it in any other way. The Keep button is deprecated and we
discourage the usage.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Gonzalo Odiard <gonzalo@laptop.org>
The keep button has led to a false assumption: learners
confused it with a traditional save button and not recognising
it as what it is, an option to duplicate an entry.
Eben suggested to remove the button [1] and others have since
then. The functionality of duplicating an entry will be moved
to the entry palette in the Journal and the entry detail view [2].
We will print a warning when the KeepButton is used and remove
it after another cycle.
[1] http://lists.sugarlabs.org/archive/sugar-devel/2010-April/023439.html
[2] http://lists.sugarlabs.org/archive/sugar-devel/2011-May/031316.html
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
This patch depends on the addition of the SugarTimeoutIcon
to the artwork where we set the font and background color.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Benjamin Berg <benzea@sugarlabs.org>
Our custom entry widget, which was based on sexyicon, did
support placing an icon in the entry (e.g. the magnifier icon
for the search entry) and displaying progress in the case of
the url entry in Browse. Both is supported since GTK+ 2.16 in the
gtk.Entry.
The patch does keep the old API, that is why we can land it in
0.94 no change in the shell needed only a small change in Browse,
in a second step I would like to use the upstream API e.g.
's/set_icon_from_name/set_icon_from_icon_name'.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-by: Benjamin Berg <benzea@sugarlabs.org>
I had to modify the patch slightly because of
0269f819761cfd803cf4d694298c8ab6a753e242, changes
reviewed by Sascha.
Signed-off-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-By: Simon Schampijer <simon@laptop.org>
Reviewed-by: Sascha Silbe <silbe@activitycentral.com>
The patch updates the window title when the user changes
the title in the toolbar or in the Journal.
Signed-of-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-By: Simon Schampijer <simon@laptop.org>
This incomplete feature contributes to confuse new activity authors and
slightly complicates our bundle installation logic.
The day someone finds something useful to do with the MANIFEST
specification, we can revert this patch in no time.
Signed-off-by: Bernie Innocenti <bernie@codewiz.org>
Tested-by: Bernie Innocenti <bernie@codewiz.org>
Tested-by: James Cameron <quozl@laptop.org>
Tested-by: Sascha Silbe <silbe@activitycentral.com>
Reviewed-by: Sascha Silbe <silbe@activitycentral.com>
[rebased on git master, minor style fix]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
To avoid that we try to access the datastore entry before
it has been created we need to move the creation code up,
before we do get the possible information from a shared
session. If we have a shared session we do then get the
title and icon-color information from it and adjust
accordingly.
Tested that no other operation like for example resuming
has issues with that change.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
To add the accelerator to the ToolButton the activity must have set
the 'sugar-accel-group' before. The patch does make the ToolbarButton
listen to the 'hierarchy-changed' signal and repack itself accordingly.
Since the ToolButtons of the subtoolbar do listen to 'hierarchy-changed'
as well to set the accelerator they will set it accordingly.
This fixes the accelerators for new-style-toolbar activities
like Terminal, TurtleArt and Paint, more info in #10930.
Signed-by-off: Gonzalo Odiard <gonzalo@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
In Python 2.7, shutdown_request was added to the BaseServer class and
TCPServer was adopted to use it instead of close_request in the
_handle_request_noblock() path.
GlibTCPServer must be adapted to account for this, so that
shutdown_request is not allowed to prematurely allowed to close the socket,
in the same way that close_request is ignored.
Fixes collaboration in activities that rely on this mechanism for
sharing (including ImageViewer and Read)
Signed-off-by: Daniel Drake <daniel@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
Tested-by: Gonzalo Odiard <gonzalo@laptop.org>
Tested-by: Simon Schampijer <simon@laptop.org>
The generic path does not work for the telepathy logs. As we
have removed the support for different profiles it is ok
to just use the default path.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Reviewed-By: Marco Pesenti Gritti <marco@marcopg.org>
To determine if the title has changed we use the 'focus-out'
event now instead of the 'changed' signal of the gtk.Entry. This
removes races we had before.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-By: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
When we join a shared activity we use the name of the shared activity
as the title of the activity. As we set the activity metadata
accordingly this will be visible in the activity toolbar, the Journal
and the activity frame.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
This differentiates between sugar and non sugar invitations using the
handle type. For non-sugar invitations we use the activity id to
determine which activity to join.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
This is a follow-up patch to 95b4eeec758ffa729d0dbb219b21d428115fcc74
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
This fixes a copy and paste error. A good test case is the Memorize
activity since we listen for the 'buddy-left' signal there to
update the members list inside the activity.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-By: Sascha Silbe <silbe@activitycentral.com>
As a side effect, activity start-up is now subjectively faster, maybe even
objectively (due to avoiding a resize).
Tested on XO-1 running Debian Squeeze.
Acked-By: Simon Schampijer <simon@laptop.org>
Before only the buddies that were present when closing the activity
were logged in the Journal. This patch does add another dictionary
'_joined_buddies' to keep track of the users that did join. The
'_buddies' dictionary keeps on tracking the users currently in the
activity.
Signed-off-by: Simon Schampijer <simon@laptop.org>
Acked-by: Sascha Silbe <silbe@activitycentral.com>
GTK pops down the palette before it invokes the actions on the
menu item. We need to postpone destruction of the palette until
after all signals have propagated from the menu item to the
palette owner.
Symptoms included the "View Details" palette menu item in the Journal list
view not working.
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Acked-by: Simon Schampijer <simon@laptop.org>
Remove the path parameter as it is not used anyhow. The patch
for the sugar module does make sure the method is called
correctly.
Signed-off-by: Gonzalo Odiard <gonzalo@laptop.org>
Acked-By: Simon Schampijer <simon@laptop.org>
Reviewed-By: Sascha Silbe<silbe@activitycentral.com>
Allow the toolbar_box to be removed by setting it to None, like is supported
for canvas.
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Acked-by: Simon Schampijer <simon@laptop.org>
Python 2 will never have pgettext() [1], so we need to ship our own version.
sugar.activity.i18n is the best place for that.
[1] http://bugs.python.org/issue2504#msg122482
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Acked-by: Simon Schampijer <simon@laptop.org>
sugar.bundle.bundle.Bundle users (e.g.
jarabe.model.bundleregistry.BundleRegistry._add_bundle) only expect
MalformedBundleException to be thrown, not exceptions of zipfile.
This patch solves the most severe issue in #1876: filling up the
filesystem with temporary files that won't be deleted afterwards.
Before we can consider this bug completely fixed, we still need
to do something for the remaining issues:
1) Unpacking shouldn't be attempted if there isn't a safety margin
2) System becomes unresponsive during unpacking
3) No progress indication for the operation, so users are tempted
to click multiple times
4) No error messages displayed for unpacking errors, which is a
common Sugar nuisance.
http://bugs.sugarlabs.org/ticket/1876#comment:5 offers possible
strategies for (1) and (2).
Signed-off-by: Martin Dengler <martin@martindengler.com>
Signed-off-by: Bernie Innocenti <bernie@codewiz.org>
[style fixes, adjusted description]
Signed-off-by: Sascha Silbe <silbe@activitycentral.com>
Reviewed-by: Aleksey Lim <alsroot@activitycentral.org>
Color data is stored in the Journal as a JSON dump, so we can get back
arbitrary types. XoColor already checks the color string for validity and
handles invalid strings gracefully, so it makes sense to enhance this to
invalid types as well.
Acked-by: Simon Schampijer <simon@schampijer.de>
As we do create the ActivityBundle in the config of the bundlebuilder
we can use the code from the activitybundle as well to install
the mime type.
Reviewed-By: Sascha Silbe <sascha-pgp@silbe.org>
pylint isn't smart enough to figure out the return type of Popen.communicate(),
so squelch the warning.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
"format" is the name of a parameter of the function we are wrapping, so we
shouldn't change it.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
We know that our code is correct and column should always be assigned, but it
never hurts to act defensively and guard against bugs in other pieces of the
code (or other components like GTK).
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
Caught by pylint. I wonder if we need this piece of code as nobody every
triggered it.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
This avoids the overhead from the string formatting on production systems.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
logging.exception() handles exceptions nicely for us, no need to explicitly
use traceback.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
Adapt to upstream format change.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
Tomeu prefers ' for strings, so let's use it wherever we don't have a good
reason to use ".
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
Most of the code uses """, so adjust the few deviations.
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
pep8 wants inline comments to be prefixed by at least two spaces. Let's just
move them to a separate line as that's easier to read anyway.
Acked-by: Simon Schampijer <simon@laptop.org>
CC: Aleksey Lim <alsroot@member.fsf.org>
Caught by PEP8. This is important for Sugar because the XO has a small screen
where long lines would make the code hard to understand (because you need to
constantly scroll horizontally).
Reviewed-by: James Cameron <quozl@laptop.org>
Acked-by: Simon Schampijer<simon@schampijer.de>
CC: Aleksey Lim <alsroot@member.fsf.org>
- add class NormalizedVersion to parse and compare the new activity versions
- change the bundlebuilder and activitybundle to use the new scheme, instead
of an int version we expect a string that matches the format defined in
NormalizedVersion
* sugar/presence/presenceservice.py: Specify the D-Bus interface when
calling ActivityProperties.GetActivity
* sugar/activity/main.py: Set a default for the --invite option and
make the create() D-Bus method accept a{sv} so we can pass the
boolean value.
When activities are uninstalled, profile data persists. Profile
data can consume a lot of storage space. Therefore it is necessary
to delete it when required.
Signed-off-by: Martin Abente <mabente@paraguayeduca.org>
src/sugar/activity/activity.py: If the activity is being invoked to
handle an invite, create a Client.Handler instance and share the
activity when HandleChannels is invoked.
src/sugar/activity/activityfactory.py,
src/sugar/activity/activityhandle.py,
src/sugar/activity/main.py: Add a -i switch that indicates to the
activity that it should handle the channel from an invitation.
src/sugar/presence/activity.py: Expose Activity.room_handle.
src/sugar/presence/presenceservice.py: Add get_activity_by_handle().
src/sugar/presence/util.py: Add get_account_for_connection().
* src/sugar/activity/activity.py: Get the shared activity wrapper from
sugar.presence.
* src/sugar/activity/activityfactory.py: Disable checking in the PS for
activity_id duplicates.
* src/sugar/presence/activity.py: Remove the PS dependency and query Telepathy
directly. Implemented enough to join an activity.
* src/sugar/presence/buddy.py: Remove the PS dependency and query Telepathy
directly. Implemented enough to join an activity. Added an Owner subclass of
Buddy()
* src/sugar/presence/presenceservice.py: Remove the PS dependency and query Telepathy
directly. Implemented enough to join an activity.
* src/sugar/presence/util.py: Add ConnectionManager for discovering and tracking
connections.
Adds the ErrorAlert class which is an alert inherited from
the base Alert class. This is very similar to the
ConfirmationAlert class with the difference being that it
only displays an 'Ok' button in the Alert popup.
Signed-off-by: anishmangal2002 <anishmangal2002@gmail.com>
The error we were seeing was that the '(env.get_profile_path(),
'data'))' does not exist yet at system start. Symlinks are created
there when accessing data on an external device. The directory
gets created when an activity like Browse or Terminal has been
run for the first time or you start Read for the first time by
resuming the file on the external device.
http://dev.laptop.org/ticket/10218
Some distros retain config files after removing packages, so /etc/olpc-security
might exist even if rainbow is not installed.
Signed-off-by: Sascha Silbe <sascha@silbe.org>
Most content bundles use global_name as the ID-style thing, as suggested
on the OLPC wiki. bundle_class seems undocumented and hence is not
present in any of the content bundles I have here.
Change get_bundle_id() to fall back on the global_name if no bundle_class
is set. This fixes a complete sugar crash when installing the standard
content bundles (it tried to send a None value over dbus).
Future extensions:
a) we are now able to reliably detect when an activity's main process has
exited. We should take advantage of this to give better feedback on activity
crashes; e.g. by turning the frame's activity-icon into an error icon and by
inserting the activity's log into the datastore so that it appears in the
Journal.
b) in the future, I'd like rainbow to provide more narrow interfaces for
manipulating jails; e.g. for garbage-collecting them, nuking their contents, or
for reusing them. Please consider this an invitation to show me which interfaces
would work best for Sugar.
P.S. - Thanks to Sascha Silbe for testing and for improvements to this patch.
Don't try to align the palette to the edges if the invoker is AT_CURSOR,
and always try to move the palette completely into the screen if at all
possible.
This attempts to align second-granularity wakeups across the system,
reducing CPU wakeups a little.
I considered all timeouts of duration 1 second or greater, and moved
the non-timing-critical ones to g_timeout_add_seconds(). (Actually,
I didn't find any that I felt were timing-critical)
sugar-install-bundle is broken because ActivityBundle.install() gained a
required parameter. Please can we make it optional? This also fixes one
of the problems with the activity updater.
self._cleanup does the self._file.close() anyway if the file exists.
The self._file.close() removed here is redundant, and will always
fail on file not found, which is handled in send_head by returning
None for self._file.
The drag-active property can be set to provide a highlight
when the tray control is accepting target of an ongoing drag.
We can't use the drag_highlight method because we desire to
adjust the background of the widget, rather than drawing on
top of it. Overriding these methods would work in most cases,
but in cases where GTK+ calls drag_unhighlight itself things
could break, so we use a property instead.
The software updater was deleting ~/Activities when it encountered a bad
bundle. Two separate issues: we didn't sanity check the bundle root to
ensure it wasn't '.' or '..' or something crazy like that, and our "clean
up on failure" code was deleting the install_root instead of the activity
root (!). This was a regression introduced by the fix for #7733 in
commit db2d1c42e2481d6dbc15405840ac23e46eab7318 (0.82.2).
The gtk Menu class supports 'table menus', where the menu entries can
be laid out in a grid. This is useful for horizontal menus, and for
dense menus of icons, and for wrapping the menu when it gets too long,
etc.
The Menu.attach() method is used to add entries to a menu by row and
column:
http://www.pygtk.org/docs/pygtk/class-gtkmenu.html#method-gtkmenu--attach
However, the callback in _Menu, defined in palette.py, only triggers
on do_insert(), and so the signal emission (and subsequent
palette._update_separators() call, which places the line underneath
the menu title text at top) never occurs when you add entries using
attach(). _Menu now overrides attach() and emits the proper signal.
Threads and signals don't get along too well together. Instead, use
gobject's spawn_async functionality which already has the necessary zombie-
reaping integrated into the gobject event loop.
