Both current design and new mockups use square icons. Currently this
is implemented by setting a fixed width and height in the CSS, which
does not work for the growing/shrinking items in the new sidebar.
By moving the assumption of square items into the code, the size will
adjust according to icon size, spacing, padding and font size if not
explicitly overwritten in the CSS.
https://bugzilla.gnome.org/show_bug.cgi?id=625887
The current design features a sidebar where the icon size shrinks
automatically to fit more items. Add a setSizeManually parameter
to BaseIcon, which disables sizes read from CSS and allows to use
BaseIcon.setIconSize() to support this.
https://bugzilla.gnome.org/show_bug.cgi?id=625887
All mockups now use a representation for documents/places similar to
the one used for applications. Rename AppIcon to BaseIcon and move its
code together with WellGrid out of appDisplay to stress their general
usefulness.
https://bugzilla.gnome.org/show_bug.cgi?id=625887
The actor emitted by ShellTrayManager is now ShellTrayIcon, a subclass
of ShellGtkEmbed which has several properties on it which are (or will
soon be) useful to the shell.
Part of the rearranging to use ShellTrayIcon means that we now show
the ShellEmbeddedWindow before creating its ShellGtkEmbed, which
requires a few modifications to ShellEmbeddedWindow (notably, telling
it at construct time what stage it will be drawn on, since it needs to
know that before it has a ShellGtkEmbed now).
https://bugzilla.gnome.org/show_bug.cgi?id=608869
Resolve a notification's ShellApp before posting the notification, so
that notification sources now always have an app associated (assuming
they came from an app as opposed to the command line). Index sources
by PID rather than by appName (so that, eg, multiple calls to
notify-send now show up as separate sources).
For Sources that have an associated app, use the app's icon for the
source, rather than tracking the notification icon.
This change also lets us get rid of appNameMap, since we can just use
shell_app_get_name() now.
https://bugzilla.gnome.org/show_bug.cgi?id=629090
Previously we were hiding the banner label if the title was too long,
but this causes queue_relayout() warnings. Instead, just set its
opacity to 0.
https://bugzilla.gnome.org/show_bug.cgi?id=629308
Introduce the Universal Access status indicator as designed, modeled
after the similar UI provided by g-s-d. This indicator allows the user
to change rapidly the keyboard and mouse behaviour (sticky keys, slow
keys, bounce keys, mouse keys), as well as the enabled ATs (magnifier,
screen reader, screen keyboard) and the HighContrast Gtk theme.
https://bugzilla.gnome.org/show_bug.cgi?id=624916
Adds class PopupSubMenuMenuItem, which is an item holding a submenu,
that is opened when the corresponding item is active, and is keyboard
and mouse navigable like its Gtk counterpart.
https://bugzilla.gnome.org/show_bug.cgi?id=623498
This patch allows the icon in PopupImageMenuItems to be changed at
any time, while preserving the item and all other properties (style,
signals) attached and without rebuilding the whole menu. This is useful
for images reflecting a dynamic status (e.g. cellular strength
indicator or battery level)
https://bugzilla.gnome.org/show_bug.cgi?id=623652
Modal dialogs slide back into the titlebar of the parent window when destroyed.
This looks weird if the parent window itself has been destroyed, so cancel the
effect in this case.
https://bugzilla.gnome.org/show_bug.cgi?id=629560
Extend WellGrid with an optional rowLimit parameter and an accessor
function for the grid's items, then use it to replace the custom
container used to display the application search results.
Also adjust the spacing of the search results slightly to make it
consistent with the app well.
https://bugzilla.gnome.org/show_bug.cgi?id=622446
The keyReleaseEvent handler does not take into account that a
user might release Alt while still holding shift
(the keycode becomes Meta_L/R in that case).
The result is that the switcher stays open which is unexpected.
https://bugzilla.gnome.org/show_bug.cgi?id=629368
Override the new mutter preference /apps/mutter/general/attach_modal_dialogs
to attach modal dialogs to their parent window. Animate the modal dialogs
expanding from the top of the parent window. Slowly dim the parent window
after the dialog comes up.
https://bugzilla.gnome.org/show_bug.cgi?id=612726
This patch ensures the following notifications behavior:
- Urgent notifications that have long title or banner text are auto-expanded
correctly.
- Single-line notifications that have _expandNotification() called (e.g.
because the user mouses over to them), are treated as expanded, which means
they get fully expanded if they are updated with more content and the user
can escape them.
- The position of expanded notifications is updated when they are updated.
- Notification banner is shown again on the first line if it can fully fit
there after a notification is updated, even if it was previously hidden
because the notification was expanded and the old banner did not fully fit.
- New notifications are immediately hidden if the user mouses away from them.
- If a new notification is updated while it is shown, we extend the time it
will be shown.
- If a new notification is updated while it is hiding, we stop hiding it and
show it again.
- If a summary notification is updated while it is hiding, we let it finish
hiding and show a new notification with the updated information.
Implementation details:
- Single-line notifications now have 4px bottom padding instead of 8px, which
means that their height matches the tray height, they are fully shown in the
banner mode, and don't pop out by 4px when the notification is expanded.
- Notification keeps a flag that indicates whether it is expanded, updates
its expanded look when it is updated, and emits an 'expanded' signal
indicating that its layout has possibly changed. The message tray connects
to this 'expanded' signal when it is showing a notification in the expanded
state and updates the position of the notification accordingly when this
signal is received so that the notification is fully shown. This is better
than connecting to 'notify::height' signal on the notification bin, since
it results in fewer callbacks.
https://bugzilla.gnome.org/show_bug.cgi?id=617209
The Shell is the only user of the magnifier, so there's no
reason to keep using GConf for its settings. Since it's a
session-wide tool, use a distinct schema,
org.gnome.accessibility.magnifier, stored in
/desktop/gnome/accessibility/magnifier just like before.
Put these settings in a separate schema file for clarity.
Old enums in GConf were stored as integers, we now use the
facilities provided by GSettings to save them as strings,
and convert them to integers automatically thanks to the
mapping stored in the schemas. Remove hard-coded default values,
which we can get from the schemas.
https://bugzilla.gnome.org/show_bug.cgi?id=622414
Return a DND.DragMotionResult constant from delegate _handleDragMotion
methods as well as the existing return value from the drag monitor method dragMotion.
https://bugzilla.gnome.org/show_bug.cgi?id=607821
Expanding notifications automatically just because they popped up where
the pointer is positioned distracts the user. It can also lead to a behavior
that is surprising to the user by, for example, making the user's input
switch to the notification's text entry box.
It is possible to expand a notification that popped up over the pointer
by mousing away from it and then mousing back in immediately.
https://bugzilla.gnome.org/show_bug.cgi?id=617209
We used to keep long titles ellipsized when a notification was expanded,
which was a bug. We now show them fully by line wrapping them.
Based on the original patch from Steven Van Bael.
https://bugzilla.gnome.org/show_bug.cgi?id=623970
We used 'bannerBody' flag to differentiate the case when we move the banner to
the body when the notification is expanded from the one when we don't do that
and only use the custom content set for the notification, as is the case for
Telepathy notifications. We also always cleared the content of the notification
on update when bannerBody was set to true.
Flag named 'customContent' reflects the use case for it more clearly. The
comments that accompany it were also updated and improved.
We now always add the banner text as the first element in the expanded
notification unless 'customContent' flag is set to true.
If the 'body' parameter is specified, we use it in addition to the banner
text. The earlier version of the code had a bug that resulted in the 'body'
parameter not being set only in the case when the 'bannerBody' was set to
true and the banner text had newlines in it.
https://bugzilla.gnome.org/show_bug.cgi?id=623970
The banner should not be appearing briefly when we are hiding the notification.
For that, we should only restore the opacity of the banner in popInCompleted()
when we are done hiding the notification. We do need to restore the opacity
in case the notification is updated and is shown in the banner mode again.
The banner should not be appearing briefly when we are showing the notification
in the summary mode. For that, we should not use the animation time to fade out
the banner in popOut() for summary notifications.
These two problems were particularly visible when the ANIMATION_TIME was increased.
https://bugzilla.gnome.org/show_bug.cgi?id=623970
In Clutter 1.2, clutter_event_get_keysym() ignored the state of the
Shift key. In 1.4, it does not, so we have to adjust our comparisons
accordingly.
https://bugzilla.gnome.org/show_bug.cgi?id=627782
The old calculation did not take into account the icon and the spacing
between columns. This resulted in the notifications that had a slightly
longer title/banner combination than could actually fit not being expandable.
The new calculation is done in _bannerBoxAllocate() so
that we don't need to hardcode which other elements are present.
https://bugzilla.gnome.org/show_bug.cgi?id=627985
A Source needs exactly one summary icon (which in the case of a
trayicon-based source won't even be just an image), but possibly many
notification icons, which may vary for successive notifications
(particularly in the case of NotificationDaemon notifications). So
differentiate these cases in the API.
https://bugzilla.gnome.org/show_bug.cgi?id=627303
The tray itself does not actually need them, and to make status icon
sources work correctly the NotificationDaemon will need to be tracking
its sources by two separate IDs, so the existing system won't work.
Also remove MessageTray.removeSourceByApp(), which is
NotificationDaemon-specific, and implement the functionality in
notificationDaemon.js instead.
https://bugzilla.gnome.org/show_bug.cgi?id=627303
The current search system uses the OR operator to concatenate search
terms. While results which are matched multiple times sort before
other matches, it is almost guaranteed that adding an additional term
to the search increments the number of results, which is rather
surprising.
https://bugzilla.gnome.org/show_bug.cgi?id=610955
When calling Notification.update, reuse the previous _bannerBox and
related labels (only changing the label content and relayouting), so
that the opacity set on popOut is preserved. As a consequence, updating
an opened notification no longer shows (or flicker) the banner at the
top.
https://bugzilla.gnome.org/show_bug.cgi?id=625502
The start date is shifted by a week if the day number of the month's
first day is smaller than the week start's day number. Probably the
only real world examples are months starting on a Sunday with locales
using Monday as start of week.
https://bugzilla.gnome.org/show_bug.cgi?id=625756
It is not obvious that only the icon is clickable to activate the
default action - in fact, with the area being that tiny, many don't
even know that notifications can be clicked.
Just extend the clickable area to the whole notification.
https://bugzilla.gnome.org/show_bug.cgi?id=613932
Adds class PopupSwitchMenuItem, which is a two-state switch similar
to those in system status indicators mockups. Also adds class Switch,
which represents the switch itself, in case it is useful somewhere
else.
https://bugzilla.gnome.org/show_bug.cgi?id=621880
In addition to already grabbing focus in the summary notifications,
we also want to grab focus in the new notifications when the user
hovers over them and expands them.
The notifications that expand automatically will only get focus once
the user hovers over them. The notifications that don't expand will
never grab focus because they don't need it.
Make sure that we toggle the way we grab focus when switching between
the overview and the main mode. This is necessary because, unlike
summary notifications that pop down when the user moves between the two
modes, new notifications keep being shown as long as the user hovers
over them or until they time out.
https://bugzilla.gnome.org/show_bug.cgi?id=617224
Because clicking on the summary item to have it display its notification
is a more deliberate action than hovering, we can now grab focus in that
notification. This makes chat notifications in the summary more convenient
to use because you don't need to click on the text entry there.
We pop down the notification when the user clicks anywhere outside of it,
triggers the overview, or the run dialog.
https://bugzilla.gnome.org/show_bug.cgi?id=617224
This ensures that the summary appears when we are showing a new
notification and switch to the overview at the same time.
Another effect of this patch is that the summary is shown if the
user moves the mouse to the bottom right corner while a notification
is being shown, which is ok.
This patch adds the method "removeMenu" to PopupMenuManager, to allow
for removal of menus after they're inserted. In order to do this, it
needs to store along with the menu all the relevant signal connections,
that are disconnected when the menu is removed.
Also adds a parameter "position" to "addMenu", so that menus can added
in arbitrary order (in particular to reintroduce those which were removed).
This patch is intended towards dynamic menu users, like extensions for
application lists, docks, sidebars showing recent documents or favourites,
as well as advanced system tray implementations.
https://bugzilla.gnome.org/show_bug.cgi?id=622730
In GSettings, a change notification is generated immediately from context of the _set() call.
In GConf, The "value_changed" signal is emitted whenever the server
notifies your client program that a value has changed in the database (100% NOT from context of the _set() call).
https://bugzilla.gnome.org/show_bug.cgi?id=624296
We need to only connect to 'action-invoked' on notification once
when the notification is created. We were accumulating callbacks
otherwise, which resulted, for example, in the Rhythmbox skipping
through multiple tracks when the 'Next' button was hit.
Always be sure to destroy the notification when its source is
destroyed, which is the case when an action is invoked.
Note that in the future, we might not want to destroy the source
for some notifications when an action is invoked. For example,
it makes sense to keep Rhythmbox in the tray while the music is
playing and allow the user to use the controls from the summary
notification too.
Solved by splitting the base class (PanelMenuButton) in a separate
module, ui.panelMenu, which is meant to hold also other reusable
button classes.
https://bugzilla.gnome.org/show_bug.cgi?id=621705
The spinner animation moving from left-to-right to indicate a starting
application changes its direction on reaching the end of the movement.
Instead, keep the spinner at the right of the app name until startup
has finished.
https://bugzilla.gnome.org/show_bug.cgi?id=623687
We should not be showing the cursor in the search entry in the overview
if the user's input will be appearing elsewhere, such as in the run dialog,
in the Looking Glass console, or in the notification's response field.
"text-align" allows setting the alignment of text, with respect to
other lines and allocated space, without requiring a reference to
the ClutterText (which is private for most widgets).
If not specified, all text is left-aligned.
https://bugzilla.gnome.org/show_bug.cgi?id=622447
Sometimes it is useful to print numbers in hex format - add the
appropriate specifier. Also support a minimum field width to
pad the formatted string with spaces (or zeros if the width is
prefixed with '0').
https://bugzilla.gnome.org/show_bug.cgi?id=622597
This is part of the design update for the message tray.
Source now takes an extra argument called 'title'.
All expanded message tray items are same width, which is determined by
the width of the item with the longest title, up to MAX_SOURCE_TITLE_WIDTH.
This is done so that items don't move around too much when one is expanded
and another one is collapsed.
https://bugzilla.gnome.org/show_bug.cgi?id=617224
While the contents of the left and right group in the panel are
flipped correctly, the groups themselves have to be swapped as
well. The location of the hot corner has to be corrected as well.
https://bugzilla.gnome.org/show_bug.cgi?id=622243
It has probably crossed the line to evil by a mile or so, but here
it is: a Tweener.slowDownFactor replacement used by all animations
without exception.
While at it, update Tweener to use the new setTimeScale() upstream
function instead of adjusting the timeline directly.
https://bugzilla.gnome.org/show_bug.cgi?id=622249
Use GSettings for all Shell configuration. GConf is kept to read
configuration from external programs (Metacity, Nautilus and Magnifier),
but ShellGConf is removed because it's mostly useless for the few calls
we still have. Also get rid of unused GConf code in ShellAppSystem.
A basic GConf schema is still used to override Metacity defaults and
configure Magnifier in a system-wide fashion. GConf is also used as
GSettings backend via the GSETTINGS_BACKEND environment variable.
All of this will be removed when these programs have been ported
to GSettings and able to use dconf.
GLib 2.25.9 is required. Schemas are converted to the new XML format,
and compiled at build time in data/ so that the Shell can be run from
the source tree. This also requires setting the GSETTINGS_SCHEMA_DIR
environment variable both when running installed or from source tree,
in src/gnome-shell.in and src/gnome-shell-clock-preferences.in.
https://bugzilla.gnome.org/show_bug.cgi?id=617917
The shell design says that upon launching an application,
no X window should have focus, and we should display an
animated launching indicator.
Implement this by in panel.js, keep track of the last started
application. If there isn't currently an X focus, show an animation
for the last starting application.
https://bugzilla.gnome.org/show_bug.cgi?id=598349
Canonical replaced status icons with libindicator based solutions,
which don't work in the shell environment. Force the distro-patched
versions to fall-back to upstream.
https://bugzilla.gnome.org/show_bug.cgi?id=621382
While zoomed out in drag mode, the workspaces at the left and right
of the active workspace are slightly transparent - when switching
workspaces, one of the transparent workspaces becomes the new active
workspace, but its opacity is not updated.
https://bugzilla.gnome.org/show_bug.cgi?id=620940
When in drag mode, the linear view raises a transparent actor
covering the entire stage. That way the view can handle events
like dragging items to the screen edges, but in order to not
disable any drop target not handled by itself, it duplicates
huge amounts of code from dnd.js.
Cleanup that mess by using the new drag monitoring support.
https://bugzilla.gnome.org/show_bug.cgi?id=620504
Sometimes it is desirable to be able to react to DND events that
happen outside a target actor's bounds, e.g. to implement reactive
screen edges. Add a simple interface which allows to hook into
drag motion and drop handling without jumping through ugly hoops.
https://bugzilla.gnome.org/show_bug.cgi?id=620504
The design calls for item to not appear abruptly, but not too slow
either - 100ms seems to be a good sweet spot for elements which are
supposed to appear "instantly".
Add a fade effect to the alt-tab popup and set the timings for other
fade effects to 100ms.
https://bugzilla.gnome.org/show_bug.cgi?id=621247
We query the transformed position and size of the popupMenu before actually showing
it which results into wrong positioning, when displaying it for the first time.
In case of the statusMenu this can even result into it being offscreen.
https://bugzilla.gnome.org/show_bug.cgi?id=621385
Add _st_actor_contains() in st-private for use within St, and
monkey-patch in a Clutter.Actor.contains() for use by javascript, and
then replace all the duplicate implementations with one or the other
of those.
https://bugzilla.gnome.org/show_bug.cgi?id=621197
Currently, we use 'gtk-yes', 'gtk-no' and 'gtk-media-pause' as
user status icons. It seems more logical to use the dedicated
icons from gnome-icon-theme instead.
https://bugzilla.gnome.org/show_bug.cgi?id=620374
First, we were passing an incorrect timestamp to
meta_display_focus_the_no_focus_window - fix that.
The invocation of set_focus_app to the started app there couldn't
really work, because (if the above call had worked) we'd get the
X reply *after* the started app.
What we need to untangle here is the distinction that's now made in
ShellApp between _STATE_STARTING and _STATE_RUNNING. A nice way to
start doing this is to rebase ShellWindowTracker to only be concerned
with app states. Concretely, the current "has windows implies
running" logic now lives just inside shell-app.c.
Rename the app-running-changed signal to be app-state-changed. This
will ultimately be useful so that inside the panel, we can track
the last started app.
https://bugzilla.gnome.org/show_bug.cgi?id=620899
Currently the workspaceSwitcher does not take the screen size into account
which could result into overflowing the screen.
Fix that but using Shell.GenericContainer instead of St.BoxLayout which takes
the monitor size into account when allocating.
https://bugzilla.gnome.org/show_bug.cgi?id=620404
Before adding a new entry to the history, we check that it does not
match the previous entry to cut down the number of duplicate entries.
Part of that condition is a check for a history length greater zero
to avoid an illegal negative index, with the side effect that nothing
is ever added while the history is empty.
https://bugzilla.gnome.org/show_bug.cgi?id=621123
As the design calls for smooth transitions whenever possible, use
a similar fade effect for the all-apps and more-docs menus as for
the panel menus.
https://bugzilla.gnome.org/show_bug.cgi?id=620775
When the app well doesn't contain any applications, it displays a placeholder text
to indicate that it's a drop target. When porting WellGrid from St.Bin to St.BoxLayout,
the call to set_child() to set the text was overlooked.
https://bugzilla.gnome.org/show_bug.cgi?id=617281
Currently, only the base style class is set in the constructor, the
style update according to the state of the represented app is deferred
until the actor is shown.
As existing icons are destroyed and recreated when the number of icons
changes, this behavior causes the icons of running apps to flicker if
CSS transitions are used.
Set the style directly in the constructor and update the code to not
delay style updates until the actor is shown - StWidget does defer
expensive calculations until the actor is mapped anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=620782
As the search entry captures the pointer when activated, the hover
state is not updated properly when the activation is cancelled
(either by clicking outside the entry or by hitting Escape).
Update the state manually in these cases.
https://bugzilla.gnome.org/show_bug.cgi?id=611095
Separate out the main app view into different sections based on the categories
in the desktop file. The configuration is done via gmenu and the desktop menu
specification, we set XDG_MENU_PREFIX="gs-" on startup, so that gmenu reads
gs-applications.menu, which we install.
There is no support for "submenus" - only the menus directly under
Applications will be displayed as categories.
https://bugzilla.gnome.org/show_bug.cgi?id=614131
Currently we only relayout when the screen size changes, this gets
the cases where a monitor gets added/removed but not when the primary
monitor changes.
We need to relayout on all monitor layout changes.
Remove ShellGlobal::screen-size-changed signal as it is no longer used, Gdk is
used to track changes now.
A ShellGlobal::gdk-screen property is added for this purpose.
https://bugzilla.gnome.org/show_bug.cgi?id=620377
In all the cases where we activate GNOME 2 UI, e.g. the shutdown
dialog, pop out of the overview. Otherwise, it's fairly broken
since you won't see the result of your action.
https://bugzilla.gnome.org/show_bug.cgi?id=620525
Currently we don't get any drop events for dual screen modes with a screen
at the right, and behave weird by hardcoding the screen size and 0 as drag
edges.
Fix that by spawning the drop group over the whole stage and take the primary monitors position and size into account in _handleDragOver.
https://bugzilla.gnome.org/show_bug.cgi?id=620378
This patch adds ISO week dates to the calendar. Week dates are an
often used feature in business and government offices. Can be turned
on through gconf, off by default.
https://bugzilla.gnome.org/show_bug.cgi?id=603532
When dragging windows in linear view, the workspace zooms out to
allow moving the window to other workspaces. Enable the same
behaviour for items dragged from the dash.
https://bugzilla.gnome.org/show_bug.cgi?id=618055
While zoomed out, the workspaces view's drop target spans the entire
monitor to implement reactive screen edges. When a drop event is not
handled by the view, an attempt is made to pass it on down the stack,
but it doesn't work properly. Fix it by iterating the target's parents
as well.
Also improve the code which translates dnd coordinates to target
positions.
https://bugzilla.gnome.org/show_bug.cgi?id=619203
Ignore the fullscreen flag when we are in the overview,
otherwise we might up not showing actors like the panel even though
they have visibleInOverview set to true.
https://bugzilla.gnome.org/show_bug.cgi?id=619693
Compute a frame rate for the period between:
- User sees first frame of overview animation
- User sees fully zoomed-out overview
And replace the current Frames count metrics with this. The
previous frame count metrics were actually over the period from
the start of the animation until the window labels finished
animating in; here we are careful to look at a more restricted
period.
https://bugzilla.gnome.org/show_bug.cgi?id=619521
Add events when we start preparing a frame and finish preparing
a frame. (In addition to measuring property-updating overhead, this allows
us to see the interval between finishing preparing a frame and starting
painting the frame, which is the relayout time.)
https://bugzilla.gnome.org/show_bug.cgi?id=619515
Currently, the drag and drop code assumes that on a successful drop
the target will either consume the drag actor or that it is otherwise
OK to destroy the actor.
As the drag behavior for window preview was changed, dropping a preview
on the dash now results in the preview being swallowed - to fix, add an
option to restore the actor in case of a successful drop as well.
https://bugzilla.gnome.org/show_bug.cgi?id=619203
The activities overview is not a place where we expect users to
interact with a specific application, so showing the application
menu there is misleading.
https://bugzilla.gnome.org/show_bug.cgi?id=618479
The preference dialog for the panel clock does not set its gettext
domain, so it's not translated even if though translations are
already available.
Fix by setting the domain in the GtkBuilder file and binding the
domain to the locale directory.
https://bugzilla.gnome.org/show_bug.cgi?id=618873
There are some places in the code where we use both fixed positioning
and CSS. Currently we use either a combination of ClutterGroup and StBin,
or we uses StBoxLayout with fixed positioning. Replace those with the new
StGroup container.
https://bugzilla.gnome.org/show_bug.cgi?id=613907
First, simply set the ellipsize flag on the application menu labels.
Next, rework how we lay out the panel components so that the center
box is always centered and constrains the left and right, rather
than pushing it around.
Previously, as part of making the shell not obviously explode if
one had a lot of tray icons, we allowed them to push the clock over.
Instead, go back to just failing in this case; we need to exile legacy
tray icons, not be slightly less ugly.
https://bugzilla.gnome.org/show_bug.cgi?id=592640
SwitcherList.actor is no longer a St.BoxLayout but a
GenericContainer now, so the hack in AppSwitcher._getPreferredHeight is no
longer needed (it is called with the correct width now).
https://bugzilla.gnome.org/show_bug.cgi?id=613194
Switch from having separate METRICS and METRIC_DESCRIPTIONS objects
in a perf module to a single METRICS array. This is done so the
perf module can define the units for each metric.
In addition to improving the output in the web interface, the purpose
of having units is to give some clue about how to pick from multiple
values from different runs. In particular, with the assumption that
"noise" on the system will increase run times, for time values we want
to pick the smallest values, while for "rate" values, we want to pick
the largest value.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
When SHELL_PERF_OUTPUT is set, instead of just dumping out the metrics, dump
a more complete report with:
- Event descriptions
- Metric descriptions and value
- Event log
Helper functions shell_perf_log_dump_events() and shell_perf_log_dump_log()
are added to ShellPerfLog to support this. The gnome-shell wrapper is adapted
to deal with the changed report format.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
Add gnome-shell options:
--perf-iters=ITERS"
Numbers of iterations of performance module to run
--perf-warmup
Run a dry run before performance tests
Make a successful run of a performance test return 0 not non-zero,
and handle the difference between that and a 0-exit in normal
usage (meaning replaced) in the wrapper.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
Add some basic statistics for allocated memory based on mallinfo(),
and use that to define two metrics:
usedAfterOverview: bytes used after the overview is shown once
leakedAfterOverview: additional bytes used when the overview is
shown a second time.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
We want to be able to summarize the behavior of the shell's
performance in a series of "metrics", like the latency between
clicking on the Activities button and seeing a response.
This patch adds the ability to create a script under perf/
in a special format that automates a series of actions in the
shell, writing events to the performance log, then collects
statistics as the log as replayed and turns them into a set
of metrics.
The script is then executed by running as gnome-shell
--perf=<script>.
The 'core' script which is added here will be the primary
performance measurement script that we use for shell performance
regression testing; right now it has a couple of placeholder
metrics.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
To support scheduling performance-measurement scripts that want to run
a number of actions in series, add shell_global_run_at_leisure() to run
a callback when all work is finished.
The initial implementation of this is not that accurate: we track
business in Tweener.js via new shell_global_begin_work(),
shell_global_end_work() functions, and we also handle the case
where the main loop is continually busy.
https://bugzilla.gnome.org/show_bug.cgi?id=618189
We don't need to reposition the menu every time its button is
allocated; we can just stick it in the right place when we pop it up
(which is guaranteed to not be during a layout cycle).
(This means that now we won't reposition the menu if the button
moves/resizes while the menu is already popped up, but it's not clear
that we'd want it to anyway, since that could easily result in the
user selecting the wrong item, etc.)
Also, we don't need to override the menu's width any more, so remove
that.
https://bugzilla.gnome.org/show_bug.cgi?id=619113
