Needs high contrast icon

Fixed in trunk.

Hi,

I have uploaded a snapshot of the current development status of Onboard containing a fix for this bug and other new features to our main PPA:
https://launchpad.net/~onboard/+archive/ppa

The snapshot is available for raring and for saucy.

Have a nice day.

This bug was fixed in the package onboard - 0.99.0-0ubuntu1

---------------
onboard (0.99.0-0ubuntu1) saucy; urgency=low

  * Request for sponsorship for new upstream release (LP: #1215164)
  * debian/control:
    - Add libcanberra-dev, libxkbfile-dev and libhunspell-dev to Build-Depends
    - Raise Standards-Version to 3.9.4
    - Remove python3-virtkey from Depends
    - Add libxkbfile1, libcanberra0, libhunspell-1.3-0 and iso-codes to Depends
    - Define new package onboard-data package for all architectures
    - Add onboard-data to Recommends
    - Update description of onboard package
  * debian/control_python2.in: removed
  * debian/control_python3.in: removed
  * debian/onboard.install: new file to split keyboard from the prediction data
  * debian/onboard-data.install: new file with the prediction data for onboard
  * debian/rules:
    - Remove support for python2
    - Add override_dh_auto_install target to use debian/tmp
    - Add override_dh_install target
  * debian/patches/add_defaults_for_ubuntu.patch:
    - Update add_defaults_for_ubuntu.patch with values for this release
  * Long press popup:
    - Hide long press popup when keyboard gets hidden
    - Add possibility to show arbitrary layouts in popups
    - Add character alternatives for most latin based languages
  * Add key-press feedback:
    - Size of label popup stored in gsettings-key
    - Also allow images in the touch feedback
    - Don't show feedback popup when xembedded
  * Add sound feedback:
    - New dependency: libcanberra
    - Add 'Play sound' option to the Preferences dialog
  * Add word suggestions:
    - Add tools to create language models (lm) from large corpora
    - Learn from typed text by default
    - Limit over-learning in the terminal with additional filters.
    - Use UTF-8 encoding internally (cuts memory usage)
    - Fall back to hard-coded country codes if there is no lm for current locale
    - Add initial system language models for: de_AT, de_CH, de_DE, en_AU,
      en_ES, en_GB, en_US, fr_fr, it_IT, pt_BR, pt_PT, ru_RU
    - Optimize system language models for size/memory usage.
    - Use case insensitive predictions most of the time
    - Offer spelling suggestions based on the hunspell C-API
    - Auto-add word separators
    - Add word suggestion row to all layouts
    - Make word suggestion also work when using the scanning mode
    - Doctests available for English, German, Italian, Portuguese, Russian and
      Spanish
  * Add auto-capitalization
    - Off by default because very new
    - Add corresponding option to the Preferences of Onboard
    - Independent of word suggestions
  * Add XInput as a new event source and make it the default: (LP: #905636)
    - Fall back to GTK event handling if XInput 2.2 isn't available
    - Allow switching between GTK and XInput events without restarting Onboard
    - Listen only to the XInput client pointer and its slaves
    - Make XInput target specific windows instead of root window
    - Support Multi-touch
    - Don't count scroll-wheel movement as button presses
    - Grab pointer even if click is performed by different device than movement
    - Fix pointer becoming unresponsive on wetab...