The login dialog had these issues:
- the entry was not really disabled, you could still edit text
- the sensitivity state was not reset on verification failure
- the session list was not disabled
The unlock dialog had these issues:
- "Login as another user..." was not insensitive
- redundant password char setting, overwriting the one given by the
question
modalDialog: do not launch disabled button actions
The "Sign In" button of the login dialog has its look disabled when the
entry is empty, but it can still be triggered by the Enter key.
This fixes the modal dialog so it does not trigger the action of an
insensitive button, and also means we do not need to connect to the
"activate" signal of the entry anymore.
Make the button layout public for callers to be able to have more
control over like adding custom widgets. Also, the clearButtons and
addButton methods are added as convenience for the most frequent usage.
MessageTray: fix popping up after there were modals
Open a modal dialog, try to open the message tray -> no effect, the message
tray is blocked by the dialog.
Close the dialog, try to open the message tray by pointer -> still
no effect, because the old timeout id was not cleared the first time,
so the dwell callback thinks the tray is about to open.
networkAgent: ensure auto-register is set when the component is enabled
Setting auto-register to true launches the initial registration, and will
afterwards re-launch the agent registration if NetworkManager gets restarted.
When the component is disabled we'll first disable auto-registration, and only
then request to unregister.
The top panel and message tray icons were by default a gnome foot and
are replaced by better ones. The applications icon is now using the
symbolic apps icon of the dash, and the windows icon is also improved.
MessageTray: restore opacity when expanding a notification
MessageTray tweens both opacity and y to hide or show _notificationWidget,
but only y when expanding it. This means that an existing tween to hide
the notification will continue running, clearing the notification state.
If the hiding one completes before the showing one, the onComplete handler
will throw an exception (because the notification was nullified) and
therefore break the state tracking.
Cosimo Cecchi [Thu, 1 Nov 2012 19:55:41 +0000 (15:55 -0400)]
remote-search: restructure remote search provider loading process
Instead of adding search providers to the system as we find them, wait
until we loaded information from all the directories, and then add all
providers at once.
This will be useful when we will sort the providers information
according to the sort order saved in GSettings.
The AnimatedIcon does not have an API for controlling the animation but
relies on the :visible property changes to start and stop a timeout used
to update the frame.
This has the inconvenient of having a side effect when visible is set to
true multiple times, and is not really the API expected from such
component. Also, there is a race if it is displayed before the images
are loaded: there is no child yet and thus we get this._frame = NaN
which leads to a crash.
Switch to a play/stop API instead, and add a load event callback to the
TextureCache.load_slice_image to exactly know when we can start using
the images.
Florian Müllner [Sat, 11 Aug 2012 03:53:59 +0000 (05:53 +0200)]
Add screenshield and unlock dialog to ctrl-alt-tab
As we now allow the ctrl-alt-tab popup on the lock screen, it should
be possible to navigate back from the top bar, so add the corresponding
elements to the switcher.
The original condition the property was based on was added to make
the a11y switcher available in the login screen, though it did never
work properly - after popping up the switcher, additional tab key
presses were ignored. As we are now able to filter bindings much more
selectively, we can simplify the check and drop the sessionMode property.
Florian Müllner [Sat, 11 Aug 2012 04:34:53 +0000 (06:34 +0200)]
windowManager: Use generic filter mechanism for all keybindings
With the new flexible system in place, there's no point explicitly
hardcoding some built-in keybindings; just use the generic mechanism
for everything.
Florian Müllner [Sat, 11 Aug 2012 00:53:38 +0000 (02:53 +0200)]
windowManager: Add a generic mechanism for filtering keybindings
Currently we hardcode the set of keybindings that are available in the
overview; add a generic mechanism to specify in which KeybindingModes
a keybinding should be available.
Florian Müllner [Fri, 10 Aug 2012 18:35:59 +0000 (20:35 +0200)]
main: Add optional keybindingMode parameter to pushModal()
For now we just use it to assign an identifier to modal modes in
which we want to allow some keybindings, but we don't use it for
any actual filtering; we'll start doing this shortly.
Florian Müllner [Mon, 20 Aug 2012 08:38:13 +0000 (10:38 +0200)]
windowManager: Implement keybinding_filter hook
We are currently using a hack to allow a select set of keybindings
in the overview. Implement the new MetaPlugin keybinding_filter
hook, which provides a cleaner way to achieve the same.
Remove section titles for ethernet and mobile broadband, and replace
them with device status items that recognize if multiple devices are
installed in the same section, and if so automatically disambiguate.
Simon McVittie [Thu, 15 Nov 2012 15:09:51 +0000 (15:09 +0000)]
st_widget_recompute_style: short-circuit if the theme node is the same
It appears to be somewhat common for st_widget_style_changed() to be
called when no style-relevant attributes have, in fact, changed. Now that
we cache theme nodes, we're likely to get the same theme node back from
the cache. If we do, we don't need to waste time asking whether its
geometry and painting are equal to itself: we can just note that nothing
really changed and get on with our lives.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465 Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
If you copy a theme node's paint state into itself, it should be an
inexpensive no-op. What actually happened was that we destroyed the
old paint state, re-initialized to blank, then copied the blank state
back into itself. In the process, we lost (for instance) the textures
for rounded corners.
Until I introduced the texture cache, this never actually happened,
because when st_widget_recompute_style() calls st_widget_get_theme_node(),
we'd always get a fresh theme node. Now, we get a theme node T back
from the cache, notice that paint_equal(T, T) is true, short-circuit
slightly by copying its drawing state into itself, and destroy drawing
state that we still needed.
I'm going to fix this in recompute_style() too, but as a general
principle, self-assignment ought to be harmless.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465 Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Simon McVittie [Thu, 8 Nov 2012 17:50:01 +0000 (17:50 +0000)]
Keep similar theme nodes so we don't have to recompute CSS so often
Because we calculate and cache CSS properties once per StThemeNode,
and only a certain set of attributes can affect the CSS properties,
it's advantageous for as many widgets as possible to share a single
StThemeNode. Similarly, if a widget changes state and then changes back
(e.g. gaining and losing the :hover pseudo-class), it should ideally
get its original StThemeNode back again when it returns to the old
state.
Here, I'm using the StThemeContext as the location for a cache.
StThemeNodes are currently never freed: this seems OK for Shell's usage
(a finite number of IDs, classes, pseudo-classes and types).
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465 Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Simon McVittie [Thu, 8 Nov 2012 15:41:31 +0000 (15:41 +0000)]
Use pre-split element classes and pseudo-classes to optimize matching
In my testing this cuts the longest time to dispatch(), when showing the
calendar menu for the first time, from 604 to 442 milliseconds,
while reducing additional_selector_matches_style() from 32% to 13% of
CPU time used.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=687465 Reviewed-by: Jasper St. Pierre <jstpierre@mecheye.net>
Florian Müllner [Mon, 5 Nov 2012 13:54:33 +0000 (14:54 +0100)]
main: Stop using Metacity's keybinding files
The descriptions for keybindings shared with Metacity are now
duplicated in Mutter, so only use the latter to avoid duplicate
entries in System Settings.
Neil Roberts [Mon, 12 Nov 2012 18:07:57 +0000 (19:07 +0100)]
Remove shell-screen-grabber
The screen grabber was a workaround for an extremely slow path in Mesa
when reading back pixel data from the frame buffer. It was using pixel
buffer objects by directly calling into GL to hit a fast blit path in
Intel's driver. This should no longer be necessary with the latest
Mesa because the normal read pixels path now has a fast path to just
memcpy the data. Using PBOs in that case just adds an extra
indirection because the data is read into an intermediate buffer and
then copied back out again.
We want to be able to remove the dependency on linking against libGL
directly from Gnome Shell because that will not work if Cogl is
actually using GLES. Also libGL includes GLX which means gnome-shell
ends up with a hard dependency on Xlib which hinders the goal of
getting Gnome Shell to be a Wayland compositor.
Florian Müllner [Wed, 7 Nov 2012 17:06:02 +0000 (18:06 +0100)]
polkitAgent: Expand password entry to the full width
Due to an oversight, the width of the password entry is currently
determined by the length of the message description. Fix the flags
so that the entry spans the entire width of the dialog.
Florian Müllner [Tue, 30 Oct 2012 19:15:29 +0000 (20:15 +0100)]
dash: Disable animations during overview transitions
When updating the dash, we already avoid all animations while the
overview is hidden. However, as we are using Main.queueDeferredWork(),
updates may be deferred up to ~20 seconds while the overview is hidden.
If the overview is entered before a queued update has taken place, it
will be run immediately on map - as the overview is visible by then,
this means animating any outstanding changes.
Work around this by skipping animations during overview transitions as
well.
ScreenShield: fix code to detect missing unlock dialog
If the session mode has no locking support, screenshield had code to
unlock automatically, but it did so by checking the return value of
the constructor, instead of checking if the constructor was actually
callable, so it would get a TypeError before reaching the check.
If I click on "Not listed?" in the login screen, I come to a username
field with two buttons: "Cancel" and "Sign In".
Clicking on "Sign In" doesn't actually sign me in though - it takes me
to the login entry. It would be better to rename "Sign In" to "Next" for
the username stage, therefore.
Gdm emit a signal to ask a question or a secret, but we can not know if
this is the last authentication question, hence we only use "Sign In"
for secret questions which improve the situation a lot.
When in the overview, if you move the mouse cursor over one of the
application launchers in the dash, all the unrelated windows are dimmed
both both in the window view and in the workspace view.
It helps to easily understand whether or not there are already opened
windows for this application, and where they are. It can also help in
differentiating the windows in the overview (sometimes the thumbnails
aren't precise enough to easily know which thumbnail belongs to which
application).
This has also the benefit of getting the application even if it can not
be retrieved through AppSystem, which can happen if the runtime WMClass
does not match the one of the desktop file.
This especially looked wrong with the following commits related to the
bug.
We simply hide the label when the popup is opened instead of relying
on the popup state when the hover state change.
To do this we replace the flag isMenuUp by a 'menu-state-changed' signal
on the AppWellIcon. This simplifies the dash label visibility handling
code that need additional changes for the bug.
Cosimo Cecchi [Thu, 1 Nov 2012 21:28:46 +0000 (17:28 -0400)]
remote-search: initialize the DBus proxy asynchronously
Initializing this synchronously means that we will possibly wait for the
process to be auto-activated and answering to our call.
If the process is already running it also might not answer immediately
our request, as it might be doing sync I/O.
The right thing to do is to initialize the proxy asynchronously; there
are try/catch blocks in place for when the object is not available, or
not properly initialized.
Cosimo Cecchi [Tue, 6 Nov 2012 22:32:39 +0000 (17:32 -0500)]
remote-search: don't use g_file_query_exists()
This is called in the main thread, which we should never block for
synchronous I/O.
Since the operation we're wrapping is async already, just use
g_file_query_info_async() instead.
Cosimo Cecchi [Thu, 1 Nov 2012 13:50:38 +0000 (09:50 -0400)]
remote-search: require a DesktopId field in search providers
Remote search providers install an auxiliary keyfile to specify
static information, such as the object path/bus name needed to activate
the binary. Such keyfiles also specify the application the providers
pushes results for; currently, we support two formats for application
information
- two fields, "Title" and "Icon" that specify a (translatable) title and
an icon name for display
- one field "DesktopId" that specifies the desktop file name of the
application backing the provider, which obsoletes the previous
Title/Icon syntax
Since all providers in GNOME use DesktopId now, and we need to ensure a
remote search providers is always backed by an application for future
development, this commit drops the support for the older syntax.
ScreenShield: account for motion velocity when hiding the lock screen
Switch from a ClutterDragAction to a ClutterGestureAction, that gives
us the velocity of mouse motion at each step, and use it to compute the
animation time for completing the hide gesture.
ScreenShield: blur and desaturate the screenshield background
The background is the same as the normal desktop, so we blur and
desaturate it to clearly show that it's not the normal system state.
To do so, we don't use standard ClutterEffects, to avoid the FBO
indirection. Instead, we take advantage of MetaBackgroundActor support
for GLSL code and paint the shaded background texture directly.
Ray Strode [Tue, 30 Oct 2012 19:06:44 +0000 (15:06 -0400)]
loginDialog: don't rely on PAM to ask for a username
For the "Not Listed?" case we will need to be able
to identify when the user has entered their username.
Once we have a way of tracking when the username is
entered, we can then defer showing the session list
too early, before the user can reliably pick a
session.
This username tracking will also be important for
implementing a disable-user-list configuration key.
If the config key gets toggled off at runtime, we'll
need to know if we're at a disruptive part of
the authentication process or not, so we know whether
we can can expose the user list right away, or wait
until the authentication conversation finishes.
Right now, we pass null in for an initial username,
and let the PAM machinery ask the user, which means we
have no good way of knowing when the username is entered.
This commit changes the "Not Listed?" code to ask the
user their username up front, before starting the PAM
conversation in much the same way we do if the user
picks a user from the user list.
a11y: fixing a regression with the accessible name at dash items
After some changes, the tooltip label at the dash is not available
until it is visually shown. As this is not anymore a reliable
source of accessible name, we just set the accessible name
with the string used on that label.
mobile-providers: new country-specific type to gather providers
shell_mobile_providers_parse() was returning the country information split
into a hash table with providers and a hash table with country names. This
patch merges both outputs into a single per-country object, so the parse()
method now returns a GHashTable with the following element-type:
(element-type utf8 ShellCountryMobileProvider>)
This also avoids more complex setups like returning lists inside of hash tables,
which was actually breaking either g-i or gtk-doc.
shell_mobile_providers_parse() was also modified to allow inputting the paths
of the country codes and provider list files to use. If paths are not given, the
default ones will be used. This helps us to provide test files during unit
tests.
Both the findProviderForMCCMNC() and findProviderForSid() methods are exported
out of the GSM and CDMA specific classes, and new unit tests for them are
implemented. Tests can be run manually with:
$> ./tests/run-test.sh tests/unit/mobileProviders.js
ScreenShield: don't rely on gnome-session to hide the lightbox
If we lock before the user becomes active again, gnome-session will never
change presence from IDLE, and thus we'll never hide the lightbox.
Instead, install our own idle monitor.
projects / gnome-shell.git / log