Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
This needs to be special cased, and requires a change in mate-polkit to work
|
|
|
|
Also catch a missed instance of needing to reposition. This stops most cases of a bottom panel menu changing size and ending up too low.
|
|
Mostly fixes serious mispositioning of network-manager applet (indicator mode) menus when connections were changed on last showing of menu and applet is on bottom panel and especially bottom right
|
|
Don't just check for the clock's existance and return if it has been destroyed, disconnect them entirely when the clock is destroyed
|
|
*On destroying applet disconnect signals for functions that otherwise segfaults with glib 2.53.4 or later.
*Fish: it's not always enough for a function to return immediately if applet destroyed anymore
*panel-action-button: suppress a warning
|
|
avoid deprecated:
gdk_screen_get_monitor_geometry
gdk_screen_get_monitor_at_window
gdk_screen_get_monitor_at_point
|
|
|
|
Fixes window list, workspace switcher. Window menu and show desktop no warnings or segfaults on removal when built in process
|
|
These two "changed" functions get triggered by clearing gsettings values on exit, make them return immediately when this happens. Don't return from dispose on repeat call or applet can't be added back
|
|
|
|
|
|
This works, but does leave part of it in RAM until the panel is restarted.
No change in out-of-process builds
|
|
|
|
|
|
Based on part of https://github.com/GNOME/gnome-panel/commit/f0f91805a1e39543487cfe67dfc47129e45e76f3
There was an issue where Wanda appeared as a thin line. The reason is
that calling gtk_widget_set_size_request() in a size_allocate() handler
doesn't work. So use an idle function for this.
Also, only call gtk_widget_set_size_request() if really needed.
|
|
|
|
|
|
|
|
Enable the force-redraw code unconditionally so it can function from the idle callback regardless of child->parent_relative_bg = TRUE/FALSE as it seems to now always be needed
|
|
fixes https://github.com/mate-desktop/mate-panel/issues/594
|
|
fixes https://github.com/mate-desktop/mate-panel/issues/594
|
|
fixes https://github.com/mate-desktop/mate-panel/issues/594
Note: min value is 75 in my tests for ukraine language
Reduced the margin, as it seemed to be the direct cause of a theme-independent jump.
|
|
fixes https://github.com/mate-desktop/mate-panel/issues/594
Note: in my language setting width_chars to 20 is minimum value to avoid
a jumping clock UI, if hovering over another location which use another day.
Maybe we need to increase the chosen value for other languages.
|
|
Support any menu theme set up for the panel's menus differing from the rest of the Gtk theme using the .mate-panel-menu-bar style class
|
|
Fixes https://github.com/mate-desktop/mate-panel/issues/570
taken from:
https://git.gnome.org/browse/gnome-panel/commit/?id=cc99dca
|
|
seems that function doesn't work proper, see
https://developer.gnome.org/gtk3/stable/GtkSizeGroup.html#gtk-size-group-set-ignore-hidden
|
|
Fixes https://github.com/mate-desktop/mate-panel/issues/520
|
|
seems that function doesn't work proper, see
https://developer.gnome.org/gtk3/stable/GtkSizeGroup.html#gtk-size-group-set-ignore-hidden
|
|
fixes https://github.com/mate-desktop/mate-panel/issues/571
taken from:
https://git.gnome.org/browse/gnome-panel/commit/?id=f93821cc77892c7b262866d04bd8aa26ddafa154
|
|
Although the specification states that most hosts will hide those,
several items report them as Passive all the time and thus never show
up. So for now, just show everything.
Fixes #537.
|
|
|
|
This fixes menu positioning on some system tray items when the panel
orientation changed since it was initially started.
|
|
Properly forward orientation to the items. This especially fixes SNI
items height in vertical panels, but affects several other details on
SNI items.
|
|
|
|
These changes have been accidentally removed. The previously used
handler has been added to the code again, which connects to the scroll
event of the WnckPager widget.
|
|
It should only affect the space between items, not all around, and
that padding should be part of the item itself, not be outside of it.
|
|
|
|
|
|
The SNI specification doesn't explicitly state what should happen when
trying to register again an already registered item, so it would seem
only logical to forbid it. Unfortunately, libappindicator tries
re-registering pretty often, and even falls back to System Tray if it
fails. So in practice we need to be forgiving and pretend it's OK,
or we get a duplicate System Tray item for this one.
Fixes #539.
|
|
|
|
|
|
The order was reversed when porting from the previous sorting method.
Part of #521.
|
|
This version check added to version 1.16 backport, let's use it in master too
|
|
|
|
|
|
|
|
It never was intended to be v3+ while upstream is v2+, so fix it.
|
|
|