GPII-2772: Windows system settings handler

[stegru]

This is the settings handler that lets you perform changes using the same back-end as the system settings applet (aka, immersive control panel).

The main piece of work is performed in an external helper C# program, because it has to be running as 64-bit. This is built during npm install (via provisioning/NpmInstall.ps1).

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/410/

[stegru]

default: WinRM transport: negotiate
Timed out while waiting for the machine to boot. This means that
Vagrant was unable to communicate with the guest machine within
the configured ("config.vm.boot_timeout" value) time period.

ok to test

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/411/

[stegru]

another CI related error

Fetching upstream changes from https://github.com/stegru/windows.git
 > git --version # timeout=10
 > git fetch --tags --progress https://github.com/stegru/windows.git +refs/heads/*:refs/remotes/origin/*
ERROR: Timeout after 10 minutes
ERROR: Error cloning remote repo 'origin'
hudson.plugins.git.GitException: Command "git fetch --tags --progress https://github.com/stegru/windows.git +refs/heads/*:refs/remotes/origin/*" returned status code 143

ok to test

[stegru]

ok to test

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/413/

[the-t-in-rtf]

So, the changes to the package.json to run the build steps as part of npm install are a key improvement, but now Vagrant builds are running that step twice. Unless there is a key reason to do both, I'd suggest updating this line:

https://github.com/stegru/windows/blob/GPII-2772/Vagrantfile#L45

[stegru]

I've removed the bits of Build.ps1 that are called in npm install.

NpmInstall.ps1 gets calles from npm install, this generates the dependencies for gpii-windows (the test-window and helper executable that this PR is for).

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/433/

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/438/

[stegru]

ok to test

==> default: Destroying VM and associated drives...
/opt/vagrant/embedded/gems/gems/httpclient-2.8.3/lib/httpclient/session.rb:623:in `rescue in create_socket': Connection refused - Connection refused - connect(2) for "127.0.0.1" port 5985 (127.0.0.1:5985) (Errno::ECONNREFUSED)

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/446/

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/472/

[amb26]

Ok oops, well at least we have learned clearly that it does not work. I've pinged @kaspermarkus to take a look at his Windows pull - in the meantime let's drop back to the previous release

[stegru]

I've updated the universal reference to use GPII/universal#594. It contains a preference set nyx which uses this settings handler.

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/473/

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/474/

[stegru]

Failures
 - python2 (exited 1) - python2 not installed. An error occurred during installation:
 The operation has timed out
VERBOSE: 'C:\ProgramData\chocolatey\bin\choco.exe' exited with 1

ok to test

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/475/

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/476/

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/482/

[stegru]

ok to test (big update to universal)

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/495/

[gpii-bot]

CI job passed: https://ci.gpii.net/job/windows-tests/543/

[amb26]

I assume you meant this to be bound to your branch #GPII-2772 which is the matching one from GPII/universal#594 - GPII-2717 looks to be a very stale branch

[stegru]

no wonder the test passed!

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/557/

[amb26]

Pretty weird integrity failure:

default: npm ERR! code EINTEGRITY
    default: npm ERR! sha1-0Ng2s0OqbbiMDez+rnHmJ11BQg4= integrity checksum failed when using sha1: wanted sha1-0Ng2s0OqbbiMDez+rnHmJ11BQg4= but got sha1-kZpOsrsMqUlqMV9U2uFSFQ36qa0=. (7551734 bytes)

npm's shoddy error reporting doesn't make it clear which package. But a big one ....

[amb26]

ok to test

[gpii-bot]

CI job failed: https://ci.gpii.net/job/windows-tests/558/

[amb26]

Proper failure this time, pretty worrying. It may be random, given I did verify that this worked locally:

09:03:26.245:  jq: Expected: {
    "gpii.windows.enableRegisteredAT": {
        "com.microsoft.windows.magnifier": [
            {
                "settings": {
                    "running": true
                }
            }
        ],
        "com.microsoft.windows.onscreenKeyboard": [
            {
                "settings": {
                    "running": true
                }
            }
        ]
    }
}
09:03:26.249:  jq: Actual: {
    "gpii.windows.enableRegisteredAT": {
        "com.microsoft.windows.magnifier": [
            {
                "settings": {
                    "running": true
                }
            }
        ],
        "com.microsoft.windows.onscreenKeyboard": [
            {
                "settings": {
                    "running": false
                }
            }
        ]
    }
}

[amb26]

ok to test

[stegru]

enabledRegisteredAT doesn't start the AT process directly, but via the shell (to start a 64-bit process) so there's a race between the real process starting, and the processReporter call. enabledRegisteredAT does return a promise for this, but it's not being passed back in the launch handler.

windows/gpii/node_modules/registeredAT/registeredAT.js

Line 143 in e2a4944

gpii.windows.enableRegisteredAT(registryName, enable, entry.options);

[amb26]

Terrible!