Screen Reader NVDA Japanese

Fork

• howto
• nvdajp
• nvdajpmiscdep

L10n updates for: nb_NO
From translation svn revision: 42399

Authors:
David Hole <balubathebrave@gmail.com>
Bjornar Seppola <bjornar@seppola.net>

Stats:
116 44 user_docs/nb_NO/userGuide.t2t
1 file changed, 116 insertions(+), 44 deletions(-)

L10n updates for: kmr
From translation svn revision: 42399

Authors:
unable to find language: kmr

Stats:
32 27 source/locale/kmr/characterDescriptions.dic
1 file changed, 32 insertions(+), 27 deletions(-)

L10n updates for: hr
From translation svn revision: 42399

Authors:
Hrvoje Katic <hrvojekatic@gmail.com>
Zvonimir Stanecic <9a5dsz@gozaltech.org>
Mario Percinic <mario.percinic@gmail.com>
Tea Turkovic <tturkovi@gmail.com>

Stats:
3 1 user_docs/hr/userGuide.t2t
1 file changed, 3 insertions(+), 1 deletion(-)

Composable Shell touch keyboard: support emoji panel and other features in Windows 10 Fall Creators Update and later. re #7273 (#7378)

* Composable Shell touch keyboard: support emoji panel in Windows 10 Fall Creators Update. re #7273.

In Windows 10 Fall Creators Update, people using United States English keyboard can use the keyboard to search for and enter emojis (Windows+period or Windows+semicolon). At the moment the app module has been added, with emoji candidates raising UIA element selected event.
Things to be aware of:
* When the emoji panel first opens, the first emoji isn't announced in some cases.
* When searching for emojis, intermediate search results are not announced.

* Emoji panel: take care of recent changes in Fall Creators Update. re #7273.

In Fall Creators Update (build 16299), when emoji categories change, the new category fires name change event.

* Readability and copyright year updates.

* Emoji panel: handle window open event in post-1709 builds. Re #7273.

In Windows 10 17000 and later (particularly after 17063 and later), when emoji panel first opens, window opened event is fired. This than allows NVDA to catch this and treat it as an item selected event, thereby allowing the first category to be announced.

* Emoji panel: support redesigned emoji panel in build 17666 and later. Re #7273.

In build 17666 and later, emoji panel was redesigned. Instead of refreshing emoji categories when Tab is pressed, one must now select a category before emoji refreshes. This is especially the case for People category where skin tone must be selected afterwards.
For backwards compatibility, continue to support original panel design in Version 1709 and 1803.

* Modern keyboard: introduce support for hardware input suggestions and cloud clipboard suggestions. Re #8189.

Modern keyboard is not just used for emoji panel entry: in version 1803, hardware keyboard input suggestions are lited there. In Redstone 5, cloud clipboard entries are shown through this panel. Thus add support for both.

* Modern keyboard: suppress 'Microsoft Candidate uI' message and handle touch keyboard's name change event. Re #7273.

In Version 1803 (April 2018 Update), 'Microsoft Candidate UI' is announced whenever candidates appear. As this is anoying, suppress this.
Also, reported by a user: under some circumstances, touch keyboard keys fire name change event, which results in NVDA announcing characters. Thus suppress this as well.

* Modern keyboard: support various changes introduced in build 17700 series. Re #7273.

Several changes as a result of deploying build 17700 series:
* Clipboard: NVDA will no longer announce label for clipboard candidates list.
* Emoji panel: NVDA will not respond to verbose name change events fired whenever items are selected.

* What's new/new features: added various modern input entries. Re #7273

* Modern keyboard/emoji panel: catch a weird case where IAccessible content generic client is the parent of the 'selected' item. Re #7273.

Reviewed by Mick Curran (NV Access): sometimes, a traceback that ends with 'AttributeError' on line 37 is shown, caused by the fact that sometimes IAccessible content generic client window becomes the parent of the 'selected' item (tree traversal issue). Thus catch this early by making sure the 'parent' is indeed a UIA control.

Upgrade to IUIAutomation6, and UIA objects: recognize more UIA dialogs via more class names and IsDialog property in Windows 10 Redstone 5 (#8473)

* UIA handler: add class names for various dialogs encountered on Windows 7, 8.x, 10, as well as apps. Re #8405.

Windows 10 and various apps added more dilaog class names. Rather than using a static list as part of overlay class chooser in UIA object, this list, along iwth new ones, are now available in UIA handler.
Also, added constant for UIA_IsDialogPropertyID in UIA handler, to be removed once IUIAutomation6 interface support comes to NVDA, as this property ID will be included in Redstone 5.

* NVDAObjects/UIA: detect more dilaogs using class names and IsDialog property in Windows 10 RS5. Re #8405.

Instead of looking up only three class names, look up more via:
* Additional class names found in Windows 10 and various apps.
* In RS5, IsDialog property.
This allows dialogs such as app uninstlal dialog and many others to be recognized and their contents read by NVDA.

* UIA/IsDialog: address review actions. Re #8405.

Reviewed by @LeonardDer (Babbage): simplify dialog flag via assigning this flag from exception hyandler segment rahter than setting it to False when COM error is thrown. This then allows IsDialog flag to be set whether or not exceptoin has occured.

* Upgrade to IUIAutomation6/IUIAutomationElement9 interface (based on build 17692).

* Use IsDialog property directly from IUIAutomationElement9 interface. re #8405.

* UIA: Popup is not supported as a fallback dialog class as the className is too generic. All the other classes are clearly dialogs from the name.

* Update What's new

Fix name mangling conflicts for the script decorator system (PR #8464)

Fixes #8463

Summary of the issue:
When using multiple global plugins with decorated scripts, the decorated script gestures were added to the GlobalPlugin base class.

Description of how this pull request fixes the issue:
When a class already has a __gestures dictionary, copy that dictionary before processing the decorated scripts. This makes sure that we do not touch the dictionary of the base class. We also only set the __gestures attribute on a class when it contains items.

Add an appModule for zoom (PR #8491)

fixes #7754

appveyor: manually grab pyscreeze 0.1.13 as latest release is completely broken (#8519)

Update espeak to commit e7e59f9 (#7846)

* Update eSpeak-ng to 1.49.3-dev

eSpeak-ng submodule is now at: 8a61f0d2108df289d713e0d74273dbfebbf23776

* Don't delete emoji files for espeak

espeak no longer freezes with emoji files present in the known cases, so
reintroduce them. This undoes the mitigation in #7810.

* eSpeak synthDriver: provide text to eSpeak as utf8 as otherwise adjacent emojis will not be combined.

* Update to latest eSpeak master. This addresses the pronunciation of words like return and selected. as espeak-ng/espeak-ng@223ff9c has been reverted.

* Upgrade to latest eSpeak master

* Update eSpeak to latest master. Most notably some improvements to Russian.

* Update espeak to latest master. Fixes #8231

* Update eSpeak to latest master

* Update to latest eSpeak master

* Minhook sconscript: correct case of hde source files as newer windows 10 builds seem to now be case-sensitive.

* Upgrade to latest espeak master

* Update What's new and readme.

Fix getting offsets from points in edit controls when offset numbers exceed 16 bits (#8397)

* Fix _getOffsetFromPoint for edit TextInfo

* Review action

* Add additional comments to make things clearer, hopefully

* Update what's new

Improve reporting keyboard layout changes (#8401)

* Improve reporting of keyboard layout changes.
In particular, fixes the problem of reporting the same keyboard layout over and over, alternating between a message with and without language name.

* Don't report if there is no change, typically on TSF profile (de)activation.
* Query the keyboard layout when NVDA starts instead of starting without a value. The previous implementation caused the first layout change to always be announced, even if the old and new layouts were identical.
* Don't allow the input method name (two placeholders separated by a hyphen) to be translatable.
* Fix a log message.
* Improve comments.
* Add copyright header.

* Remove XP code for TSF from NVDAHelper.dll.

* Update copyright header.

* Use the LOWORD macro to get a lo-word.

* Review actions.

* Review actions.

* Update what's new

UIA: assume Windows 7 is in use and remove constants that are now part of current interface (#8422)

* UIA handler: just assume UIA is available, but keep the UIA enabled key. Re #8415.

UIA support in NVDA assumes Windows 7 or later. Because NVDA supports Windows 7 (SP1) and later, just assume UIA is available. However, for cases where UIA should be turned off, turn this on if UIA enabled key (from config) is on.

* Config: remove minWinVersion key for UIA. Re #8415.

As NVDA supports Windows 7 (SP1) and later, there's no need to check if min Windows version is 7 anymore, thus the coresponding key in Config manager/UIA section is no longer applicable.

* UIA constants: remove constants added in current interface that ships with NVDA. Re #8416.

In the old days, when UIA handler supported Windows 7 and 8.x (IUIA1 through IUIA3), there were constants that were not part of the interface file that came with NVDA. now that iUIA5 is supported (and in the future, IUIA6), it makes no sense to continue to include constants that are now part of the interface itself. Thus constants included in latest UIA interface are no longer part of NVDA's own UIA handler module (not all are gone, however).

* Update what's new

Upgrade to wxPython 4.0.3 (Phoenix) (#7104)

* wxPython4: initial foundations with wx.adv.TaskbarIcon and others. re #7077.

One of the biggest benefits of using wxPython 4 is easier route for upgrading to Python 3.x. Due to changes made in wxPython 4, NvDA source code (at least wx routines) must be modified. This provides some foundations, namely wx.adv.TaskbarIcon and replacing wx.SpinCtrlNameStr with a string literal.

* wxPython 4/Python Console: use wx.TextCtrl.WriteText instead of wx.TextCtrl.write. re #7077.

wx.TextCtrl.write is no more - wx.TextCtrl.WriteText should be used instead.

* wxPython 4/Python Console: use AppendText instead of WriteText for cursor tracking. re #7077.

If wx.TextCtrl.WriteText is used and if the cursor is located at the top of the output window for the console, newly typed text will be inserted instead of being appended, thus wx.TextCtrl.AppendText will be used.

* wxPython 4/various dialogs: wx.CENTER_ON_SCREEN is no more, use wx.Center instead. re #7077.

Somehow, documentation says wx.CENTRE_ON_SCREEN is included but it isn't there at runtime. Instead, wx.Center is used (does not follow variable naming convention though), so using it for now.

* wxPython 4/Welcome dialog: wx.NORMAL cannot be used as a font family anymore, replaced with wx.FONTFAMILY_DEFAULT. re #7077

* wxPython 4/cursor manager/Find dialog: use sizer.Add instead of sizer.AddSizer. re #7077.

* wxPython 4/Launcher: spell out GridSizer gaps, prevents overload error. re #7077.

In wxPython 4, when instantiating wx.GridSizer, horizontal and vertical gaps in pixels must be specified.

* wxPython 4/speech viewer: few tweaks to constructor, save position before the dialog is destroyed. re #7077.

Due to changes to how dialog sizes are constructed, keyword arguments cannot be used.
Also, if Window.Destroy is called, somehow text controls and what not are gone, so save speech viewer position before destroying the window.

* wxPython 4/Input gestures dialog: when Cancel button is clicked, make sure to check if the treeview is alive. re #7077.

Somehow, when Cancel button is licked, treeview is removed first, yet item selection event is run. Make sure to catch this.

* Readme: mention new wxPython and Six compatibility layer requirements. re #7077

* wx.Menu: AppendMenu is deprecated, use Append instead. re #7077

* Add-ons Manager: make sure to nullify dialog instance flag and prevent issues when installing more than one add-on remotely. re #7077.

Here 'remotely' refers to letting people install add-ons via Windows Explorer and other means. Without nullifying instance flag when the add-ons manager closes, instance flag will be kept, which causes wxPython to think (and correctly) that the dialog is still active, resulting in runtime error on add-on list control being thrown.

* gui/Config profiles dialog/wxPython 4: nullify instance flag. re #7077.

Just like add-ons manager, if the instance flag for config profiles dialog isn't nullified, wxWidgets will think that the dialog is active when it is gone. Thus nullify it in two places: when the actual dialog closes and when a new profile is activated right away (for the latter, it is parent.Destroy).

* Python console: revert to wx.TextCtrl.write function. re #7077.

Due to request from a tester, wxPython 4 developer has decided to restore wx.TextCtrl.write. This means the AppendText workaround in place is no longer needed.

* Settings/voice sliders: replace wx.WXK_PRIOR and wx.WXK_NEXT with wx.WXK_pAGEUP and wx.WXK_PAGEDOWN, respectively. re #7077.

wx.WXK_PRIOR and wx.WXK_NEXT are no more, replaced by direct definition of page up and page down keys. This fixes the problem where sliders couldn't be changed in Voice Settings dialog.

* core.CorePump: block this timer from firing recursively. This could never happen before wxPython 4, but now it is possible.

* miscDeps now contains wxPython 4.

* Core: update copyright years, corrected issue number

* Comment spelling fix

* wxPython 4.0.1 in miscDeps

* gui: add a NonRe-entrant timer to replace use of wx.Timer and wx.PyTimer.

* Replace usage of wx.PyTimer with gui.NonReentrantTimer for all core code in NVDA. This does not change any braille display drivers.

* Clarify comment

* Clarify docstring

* Spelling

* Readme: mention wxPython 4.0.1

* wxPython 4: use wx.ID_ANY instead of NewID function. Re #8121.

Becasue NewID may raise assertion error, use wx.ID_ANY instead.

* GUI/SettingsDialogs: remove extra space at ends of lines

* Settings screen: wx.Center again.

* add a new submodule for wxPython dep

* Use `CenterOnScreen()` rather than `Center`

Use the explicit function rather than `Center()` with the CENTER_ON_SCREEN constant.
Also supply CENTER_ON_SCREEN for addons which may still rely on it.

* Define both missing constants

Despite being referred to in the documentation for wxPython,
CENTER_ON_SCREEN and CENTRE_ON_SCREEN constants are missing.
Here we redefine them to stop errors in addons. Addons should migrate to
using the explicit functions CenterOnParent or CenterOnScreen.

* Update copyright years and headers.

* Settings dialog: clear instances set when all settings instances are gone. Re #7077.

Previously, when closing settings dialog, instances would be cleared. It turns out that might not be the case with wxPython 4, so manually clear instances set if the only thing remaining is the SettingsDialog object.

* wxPython 4: revert instances checking in SettingsDialog thanks to mitigation about circular references. Re #7077.

* Dictionary dilaog: use wx.NewId for add, edit, and remove buttons. Re #7077.

Somehow, using wx.ID_ANY for speech dictionary buttons causes the buttons to not work. This is because the button is added directly to the helper sizer. Thus use NewID instead.

* Get rid of unnecessary ID usage

* Log viewer: wx.SAVE -> wx.FD_SAVE, wx.OVERWRITE_PROMPT -> wx.FD_OVERWRITE_PROMPT. Re #7077.

Yet another change of attribute names: wx.SAVE is now wx.FD_SAVE, wx.OVERWRITE_PROMPT is fwx.FD_OVERWRITE_PROMPT. This resolves an issue where Log Viewer does not let users save logs (issue #8385).

* Log viewer: add copyright headers.

* wxPython4/settings dialogs: deleteWindows -> delete_windows in sizer.Clear method. Re #7077

* What's new: add wxPython 4.0.1 entry in changes for developers section. Re #7077

* Upgrade to wxPython 4.0.3

* Touch intewraction/touch typing mode checkbox: no more wx.NewId function for this checkbox just like others. re #7077.

Somehow, for touch interaction/touch typing mode checkbox, wx.NewId function was called, whereas other checkboxes do not do this. Thus make it consistent with other fellows by not calling this function in this particular checkbox.

* Settings panels: use wx.NewIdRef function to retrieve global ID's for panels. Re #7077.

Reported by several people and subsequently confirmed: in wxPython 4.0.0 to 4.0.2, there exists a wrap-around bug in wx.NewId function that prevents programs calling this function from working normally after the ID's are exhausted. This is now fixed in wxPython 4.0.3 via wx.NewIdRef function. Thus use this function to keep an eye on panel ID's, used for announcing new categories as they become visible in multi-category NVDA Settings screen.

* Readme: mention Python 2.7.15 and wxPython 4.0.3 dependencies.

* What's new: wxPython 4.0.1 -> 4.0.3. Re #7077.

[skip ci]

Several fixes to the ALVA driver (#8230)

* Work around malformed feature reports sent by the ALVA BC680 when initializing

* Fix looking up modifier gestures

* Distinguish between left and right smart pad and thumb keys

* Force disable keys that trigger internal functions

* Fix assignments to kb:alt

* Review actions

* Swap date time and windows command assignments

* Revert force disabling of keys, as it does not work and this should be handled by the dvice anyway

* Update changes.t2t

Microsoft Store edition: detect if this is just a package or not (look for error 15700). Re #8362. (#8374)

kernel32.dll::GetCurrentPackageFullName returns error 15700 (not a package) when a process is run outside an appx package. Thus catch this instead of 0 (error success). This then allows NvDA appx to run correctly.

Support aria-roledescription in Edge. (#8449)

* Support aria-roledescription in Edge.

* Braille: honor roleText in browseMode.

* BrowseMode: only explicitly include the name of a landmark if it is not going to be already presented as part of the field.

* Edge: mark controlFields as block if if they have a particular role (grouping, paragraph or section currently). This ensures that they are appropriately presented as containers when roledescription is used.

* Update What's new

* Address review actions.

Add automatic background braille display detection (#7741)

* Implement braille display auto detection. This new background mechanism can be activated by choosing the Automatic option from NVDA's braille display settings. See #1271 for an in depth discussion of funcionality.

* brailliantB, use generic implementation for USB serial devices
* Use generic check function for braille display drivers supporting bdDetect
* Make auto detection the default
* Set a timeout for joining the bgThread when auto detection was on
* Poll for bluetooth devices on app switches
* Support bluetooth HID in bdDetect

* Use a separate thread for background scanning.

* Make the bdDetect thread a daemon thread

* Disable auto detection within the unit test framework. If we don't do this, detection will occur and unit tests will fail.

* in braille.handler.handleGainFocus, check whether the focused object has an active tree interceptor. If so, focus the tree interceptor instead

* Revert the use a separate thread for background scanning and make sure recursion does not occur when using an APC

This reverts commit 5b97f3952348fe80db8ab06b1e9a008c105d48d0.

* Created Detector._scanQueuedSafe which wraps changing the state of _scanQueued within a lock

* Fix malformed VID and PID for Brailliant, add an extra check

* NO longer filter serial ports for Brailliant

* Updated changes.t2t

Merge commit 'fetch_head' into betajp

System test with robot framework (#8481)

* prototype sysTest

Requires:
robotFramework - the system test framework
pyautogui - for controlling keyboard/mouse

To install:
pip install robotframework
pip install pyautogui

* WIP include flaky way of exiting

Ensure this is run from a cmd prompt. Other shells may cause issues,
for instance, NVDA does not get correct focus when run from cygwin.

* WIP System testing

Remote server implemenation and external abstracted library.

* First try at including system tests in appveyor.

* Disable pyAutoGUI's failsafe check as appeyor moves the mouse???

* initial.robot: break robot to temporarily bipass freeze.

* appveyor.yaml: run robot in powershell so exit code is ignored for now.

* appveyor.yaml: try again to make system tests return error code 0

* appveyor.yaml: try again to make system tests return error code 0

* appveyor.yaml: fix syntax of uplodFile.

* system tests: tell appVeyor they are junit formatted?

* appVeyor: still upload test results on failed builds.

* Upload unitTest reports to appViewer as well.

* Investigate freeze in robot test.

* add some logging to nvdaRobotLib

* More logging

* more logging again

* sleep before connecting to server.

* more and more logging

* Fix logging...

* reimplement quit nvda test inside initial.robot.

* A try at moving everying to robot language to fix freezes.

* Launch NVDA from robot with py for appveyor.

* copy the globalPlugin.

* Temporarily start rdp for appveyor build

* bump

* bump 2

* bump 3

* bump 4

* Slightly re-write now with setup and teardown, and insert q for quit check and check foreground name is Exit NVDA etc.

* nvdaRobotLib: use a specific NVDA profile dir.

* NVDA Spy for robot must use a daemon thread for now to stop further NVDAs crashing.

* Robot tests set NVDA not to show the welcome dialog

* allow asserting on last speech

* Remove extra tab characters

* Copy in the testSpy globalPlugin for each test.

* Remove sleep

The sleep was required because importing the remote library suceeds
regardless of whether it is possible to execute a keyword. We now test
the connection before we load the library.

* Some cleanup of connection/disconnection code

* Extract code that can be reused and add better documentation

* cleanup systemTest dir

Allow use of different settings files and some cleanup of system test directories

* Assert text on welcome dialog passes

* Copy nvda log for each test

* Try to get systemtest artefacts from appveyor

* Try manually pushing artifacts

* Try recursing over files and pushing manually

* Fix issue when nvdaProfile folders do not exist

* system tests: don't show usage stats dialog in NVDA

* appveyor: install NVDA on the build system, and run the system tests against the installed copy.

* appveyor: install nvda with start-process and wait up to 3 minutes for it to complete, failing on timeout.

* fix appveyor errors

* Temprarily include robotremoteserver in NVDA for system tests. Must figure out a better solution.

* Breakup up appveyor test code into separate powershell code chunks to better show build progress.

* system tests: use silence synth driver.

* system test for exit dialog: just sleep instead of varifying dialog.

* system tests: don't wait for exit dialog, rather just confirm that the NVDA process definitlly exited.

* appveyor: move NVDA installation to before_test section.

* ensure quit system test: exit from the NVDA menu.

* system tests: try alt tabbing to the exit dialog???

* More debugging for system tests.

* IAccessibleHandler: more debugging.

* IAccessibleHandler: more debugging.

* eventHandler: more debugging.

* Up max foreground defer count to 10.

* Up foreground defers to 100 and log them.

* IAccessibleHandler: drop foreground defers back down and allow welcome dialog in quit from keyboard system test.

* system tests for quit with keyboard: watch for and close welcome dialog first.

* quit with keyboard system test: wait longer for NVDA to exit.

* Restore IAccessibleHandler and eventHandler.

* Not able to get an index returned from get_index_of_speech

* tests passing

* move spy lib logic into spy

* Install systemTestSpy as a package on setup

* Cleanup files, add copyright headers

* Use _ to prevent imports from leaking as keywords

* Move non NVDA specific helpers into a new library

* Create self contained package for system test spy

The system test spy now copies in library dependencies that are not
provided by a NVDA install. This is so that the system tests can be run
on appVeyor against an installed version of NVDA

* Zip system test output before pushing to artifacts

* try adding robotlibraries directory to the path

* Fix case of file name

This should resolve the appveyor system test failures

* further tidy up

* Introduce suite name

* fix appveyor system tests

testOutput directories need to be created before writting files to them!

* Stop double nvda install log artifact

* set robot library path

* reduce suite verbosity on appveyor

* Revert to running from repo root

Running from tests/system/ did not reduce the number of suites mentioned in Robots reports

* Clean up some of the appveyor code

* clean up some names and docs for tests

* Add scons command to run system tests

* ensure dir is created for std out/err logs

* Use dummy speech synth to get speech

This captures all NVDA manipulation of speech before it is given to
synthesizers.

* wait longer before sending input

* Stop update dialog from showing on startup

* simplify the appveyor pip installs

* Update What's new

{liblouis3.6.0 brailletables} (#8446)

* added czech eight dots computer table

* added hungarrian braille table

* update braille tables file

* review actions

* review actions,
note, if the unit tests pass, this can be merged into master

* review action, change k to K in Central Kurdish

* update changes file for #8446,
based on latest master

* revert the documentation edit, as thic can cause conflicts. This can be merged straightforward, as the only master branch is updated.

* Update changes.t2t

* Update changes.t2t

Update translations.

From translation svn revision: 42378

L10n updates for: zh_CN
From translation svn revision: 42378

Authors:
vgjh2005@gmail.com
jiangtiandao901647@gmail.com
manchen0528@gmail.com
sysking1024@gmail.com
dingpengyu06@gmail.com

Stats:
10 12 source/locale/zh_CN/symbols.dic
4 12 user_docs/zh_CN/userGuide.t2t
2 files changed, 14 insertions(+), 24 deletions(-)

L10n updates for: vi
From translation svn revision: 42378

Authors:
Dang Hoai Phuc <danghoaiphuc@gmail.com>, Nguyen Van Dung <dungnv1984@gmail.com>

Stats:
7 7 user_docs/vi/changes.t2t
3 2 user_docs/vi/userGuide.t2t
2 files changed, 10 insertions(+), 9 deletions(-)

L10n updates for: kmr
From translation svn revision: 42378

Authors:
unable to find language: kmr

Stats:
40 194 source/locale/kmr/LC_MESSAGES/nvda.po
1 file changed, 40 insertions(+), 194 deletions(-)

L10n updates for: de
From translation svn revision: 42378

Authors:
Bernd Dorer <bdorer@mailbox.org>
David Parduhn <xkill85@gmx.net>
Rene Linke <rene.linke@hamburg.de>
Adriani Botez <adriani.botez@gmail.com>
Karl Eick <hozosch@web.de>
Robert Hänggi <aarjay.robert@gmail.com>

Stats:
25 179 source/locale/de/LC_MESSAGES/nvda.po
38 38 user_docs/de/userGuide.t2t
2 files changed, 63 insertions(+), 217 deletions(-)

Merge branch 'master' into beta

work around appveyor stuck

modified version for betajp noncert

Update translations.

From translation svn revision: 42340