Age | Commit message (Collapse) | Author |
|
Fixes [YOCTO #13319]
Signed-off-by: Mingde (Matthew) Zeng <matthew.zeng@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The soft keyboard was displaying incorrectly on 64bit machines
[YOCTO #9254]
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
XML_GetCurrentLineNumber() returns unsigned long as clang is telling us
here
src/config-parser.c:603:6: warning: format specifies type 'int' but the
argument has type 'XML_Size' (aka 'unsigned long') [-Wformat]
XML_GetCurrentLineNumber(p),
^~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
sequence"
No longer needed, as commit 03c704f84963d2aefeede41edd76eb36caee96f6 provides
a better solution.
This reverts commit 0899b0260fe27ffa95ac200835745f32d7bb44bb.
Signed-off-by: Tomas Frydrych <tomas@tomtain.com>
|
|
When the kbd window maps, it is possible that it is not using the current
backing pixmap. The preceeding commit has was attempting to address this
problem by forcing backbuffer setup after an unmap event, but this solution
fails when the backbuffer pixmap is not attached on the initial map, which
seems to happen with more recent versions of gtk.
Signed-off-by: Tomas Frydrych <tomas@tomtain.com>
|
|
When the widget is going from an unmapped to a mapped state, if it was
previously mapped at the same size, it is possible for the backend to end up
attached to a stale drawable. So when the widget gets unmapped, set a flag
to force fresh backbuffer set up on any future map at an identical size.
(If the new map happens at a different size, this triggers complete relayout,
which included backbuffer set up.)
Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
|
|
|
|
|
|
If a key has a modifier then on release we have to redraw the whole keyboard
because on release the modifier is reset, so the kbd layout changes. The
Caps modifier is an exception, since it stays in place until it is explicitely
cancelled, but we were still scheduling a full redraw; if the user drags a
finger across the kbd and the HW is not fast enough to keep up with the
redraws, this becomes rather obvious.
|
|
Such behaviour is not nice to the calling application.
|
|
Normally the locale gets determined automatically from the environment, but
there are times when an application might want to force a different locale
(e.g., en_US).
|
|
Only check for long press for the unshifted -> shifted transition, for
shifted -> unshifted do just normal processing
Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
|
|
Patch by Wim Crols
|
|
We need some custom attributes here, so make our own <include/> tag rather
then implementing XInclude machinery
Attributes:
* 'file' - what to include; path to be relative to layout directory
* 'auto-locale': yes|no, default yes -- when 'yes', locale specific version
of the include is loaded, same as is done for layouts themselves.
|
|
|
|
If the KeySym is in the upper case range, force the SHIFT flag, if it is
in the lower case range, forcibly clear the flag.
|
|
If a kbd layout contains a capital letter that is not part of the server
keyboard map, we need to propagate associated shift state to libfakey
to ensure that a synthetic event is generated for the Shift key; without
this the keyevent will not have the keyboard state set correctly, which
might result in the lower case peer symbol emerging in the application.
|
|
The gdk foreign window API is deprecated and we are getting some
warning because Gdk can properly handle foreign window.
|
|
* Added x,y parameters to mb_kbd_ui_resize() to allow combinging move
with resize; made public
* split out backbuffer resizing from mb_kbd_ui_resize()
* When style changes, resize backbuffer to ensure that it is still
attached to the window (i.e., GtkWidget::set_style messes with the
window background.
* In size_allocate(), use mb_kbd_ui_resize() rather then resizing the
Gdk window
* Do not process ConfigureNotifies for UI window -- this is our OR
window only we manipulate
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The _allocate() function calculates the minimal size of the kbd based
on the requested font size; as such it should only be run once when we
create the widget (running it during resize was causing the stored
base font size and dimensions to be out of sync, leading to wrong
calculation of real font size).
git-svn-id: https://repository.vernier.com:23845/repos/trunk/vernier/sources/matchbox-keyboard@9832 37671d33-5a16-0410-807e-c55260944247
|
|
|
|
Make long press on the shift key behave like a caps lock
(Only implemented for the GtkWidget at the moment; the only reason for
this is I can't be bothered to implement a timeout function ...)
|
|
* Split the code into a thin program wrapper and a library.
* Implemented a native Gtk widget, MbGtkKeyboard using the library
matchbox-keyboard: pc file
|
|
Due to a bug in GtkSocket (at least in 2.14), we only get mapped if
at the point of requesting to be mapped we have a pending size change;
if not, the allocation process that contains the XMapWindow() call
shortcircuits. Work around this by mapping ourselves once we are
embedded.
|
|
|
|
* Adds a popup above the key on press (except for modifiers)
* With the exception of Backspace, keys only generate key events on
release (this means no autorepeat)
|
|
* ignore resize requests with dimension <= 10 (happens on initial
embedding, and is followed by a proper request later)
* reload font when either dimension changes, rather than both
|
|
To allow to specify initial size of the window
|
|
|
|
* Refactor things, so that instead of various globals we have per-
widget data
* Implement object tear down
|
|
Nothing like home-grown fruit an vegies ;-)
|
|
TODO: the colours should be easier to customize
|
|
|
|
mb_kbd_ui_resize() was exiting early if the height assigned to the
kbd was smaller than the minimal height the UI requested; unfortunately
this leaves the keyboard window with wrongly sized backing pixmap --
we must respect the size assigned to us, whether we like it or not.
|
|
Lowered the height restriction from 1/3 of screen height to 2/5 of
screen height.
|
|
|
|
|
|
* src/matchbox-keyboard-image.c:
Include render headers.
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@2075 b067294f-1dea-0310-9683-c47a78595994
|
|
wide screens
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1910 b067294f-1dea-0310-9683-c47a78595994
|
|
* layouts/keyboard.xml:
Add UTF8 arrow keys (thanks to pippin)
* src/matchbox-keyboard-ui.c: (mb_kbd_ui_event_loop):
Disable gestures at least fo now.
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1739 b067294f-1dea-0310-9683-c47a78595994
|
|
* src/matchbox-keyboard.h:
* src/matchbox-keyboard-remote.c:
* src/matchbox-keyboard-remote.h:
* src/Makefile.am:
Move the remote operations into a separate header.
* gtk-im/im-protocol.c:
* gtk-im/im-context.c:
* gtk-im/Makefile.am:
* gtk-im/im-protocol.h:
* applet/applet.c:
* applet/Makefile.am:
Use src/matchbox-keyboard-remote.h instead of a copy of the enum,
to avoid stupid bugs. Thanks to Daniel Willmann for noticing.
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1736 b067294f-1dea-0310-9683-c47a78595994
|
|
* src/matchbox-keyboard.h:
* src/matchbox-keyboard-ui.c:
* src/matchbox-keyboard-remote.c:
* gtk-im/im-protocol.h:
Add a TOGGLE action to the RPC.
* applet/applet.c:
Instead of using a toggle button use an event box, and sent the
toggle event.
Based on patches by Stefan Schmidt and Daniel Willmann.
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1732 b067294f-1dea-0310-9683-c47a78595994
|
|
* src/matchbox-keyboard.h:
* src/matchbox-keyboard-remote.c:
Change mb_kbd_remote_process_xevents to return
MBKeyboardRemoteOperation instead of performing the operation.
Also Declare mb_kbd_ui_limit_orientation.
* src/matchbox-keyboard-ui.c:
Action the operation reutrned returned by process_xevents, and if
we receive a Hide command wait for 100ms before hiding the
keyboard, so that if we immediately receive a Show the keyboard
doesn't flicker.
Based on a patch by Daniel Willmann <daniel@totalueberwachung.de>.
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1687 b067294f-1dea-0310-9683-c47a78595994
|
|
* src/matchbox-keyboard-key.c:
Dont set uneeded shift flag in libfakekey (Daniel Willmann)
git-svn-id: https://svn.o-hand.com/repos/matchbox/trunk/matchbox-keyboard@1675 b067294f-1dea-0310-9683-c47a78595994
|