Test if running on simulator

I don't think there is any documented way to do so. This is where some sort of conditional compilation would be really nice. It seems that we could use annotations to solve this problem, but without the ability to define an annotation on the compile line, I don't really see how it would work.

You may be able to use knowledge about simulator quirks to your advantage. The problem with this is that the results could be wrong if the simulator or the device firmware fixed. As a (possibly incorrect) example, I believe Sys.println() will raise an exception if called on something that is not a Lang.String on the device, but it works just fine in the simulator. You'd have to verify this behavior, or find another similar quirk, and keep this function updated.

function guess_is_simulating() {
var result = false;
try {
Sys.println([ 1 ]); // assumes that this crashes when running on device firmware
result = true;
}
catch (e instanceof Lang.Exception) {
}
return result;
}


I don't believe you can leverage string resources for the sim devices (like fenix3_sim), but if you can, you could probably use that to your advantage as well.

A while back I found something that you might use (I just tried it on the current SDK)

Sys.getDeviceSettings().firmwareVersion

In the Sim, it's [0,0] but has the real info if running on a device. If major and minor are both zero - it's running in the sim

Yes, that sounds much better than my 'suggestion'. It seems that this should work then...

Note: This code does not work as of the ConnectIQ 2.2.0 SDK.

function guess_is_simulating()
{
var deviceSettings = Sys.getDeviceSettings();

if (deviceSettings != null) {
return (0 == deviceSettings.firmwareVersion[0] &&
0 == deviceSettings.firmwareVersion[1]);
}

return false;
}


Travis

Perfect, thanks.

Dave.

It appears that the 2.2.0 SDK now gives non-zero values for the firmwareVersion, so the above code no longer works. While it isn't as nice as a runtime check, it seems that you should be able to do conditionally compile the following code...

(:debugging) function guess_is_simulating() {
return true;
}

(:shipping) function guess_is_simulating() {
return false;
}


If you go to Windows > Preferences > ConnectIQ > Compiler and add -x debugging to the Compiler time options, the first function will be used. When you get ready to build your final version, change the compile to -x shipping.

Travis

Bummer that the firmwareVersion isn't a flag for the sim anymore, but it make sense with checks for fw/CIQ versions now in the SDK/app store so I get it.

Maybe in a future SDK, Garmin could add something like "Sys.getDeviceSettings().inSim" so runtime checks could be done. Using "-x" works, but it's something that's easy to forget and leave in the incorrect state.

TL;DR Test for presence of TZ/DST bug that only exists only real devices [1/2] EDIT: this idea is bad, don't use it

I have a way to test for the sim, but it's terrible. There are 2 prerequisites:

1) the simulator is running on a computer whose configured timezone/location observes DST in July and standard time in December, or vice versa. (Specifically, the time zone offset in December must be different than the time zone offset in July). The Garmin device must also fulfill the same condition.

For example, the time zone offset in New York is -5 hours in December and -4 hours in July, so this test will work if you are there.

Unfortunately, if you live in most of Saskatchewan, most of Arizona, Hawaii, mainland China, Hong Kong, etc., where it's either DST all year long or standard time all year long, this trick is not for you :/.

TL;DR Test for presence of TZ/DST bug that only exists only real devices [2/2] EDIT: this idea is bad, don't use it

2) Garmin never fixes the following bug: Gregorian.info() handles DST incorrectly for dates other than now/today (works in sim, not on device). [This applies to passing in a Moment to Gregorian.info(), which obtains information for the device's local time zone]

I reported it 3 years ago, I'm sure it's existed forever, and it still exists on an FR955 running recently released beta firmware. It looks like it was fixed (or never existed) when a LocalMoment is passed into Gregorian.info(), which obtains information for the time zone/location specified by the LocalMoment. The lack of this bug for LocalMoment unfortunately means that this test does rely on the Garmin device's actual location.

function isSimulator() as Boolean {
    // ‼️‼️‼️Assumes both the simulator and device are in a location which observes DST
    // in July and standard time in December, or vice versa.
    // (Specifically, the time zone offset in December must be different than the time zone offset in July).
    return !haveDstBug();
}

function haveDstBug() as Boolean {
    // Test for the presence of the following bug, which should exist only on a real device and not the simulator
    // https://forums.garmin.com/developer/connect-iq/i/bug-reports/gregorian-info-handles-dst-incorrectly-for-dates-other-than-now-today-works-in-sim-not-on-device
    // Gregorian.info() handles DST incorrectly for dates other than now/today (works in sim, not on device)
    // In the sim, if I create a moment for some other date outside of the current DST/non-DST period, Gregorian.info() outputs local time correctly (with the correct adjustment/non-adjustment for DST).
    // On a real device, the same test will produce output which acts as if the supplied date is today, for the purpose of DST adjustment.

    // THIS TEST ASSUMES THE COMPUTER RUNNING THE SIMULATOR IS CONFIGURED FOR A LOCATION/TIMEZONE
    // WHICH OBSERVES DST IN JULY AND STANDARD TIME IN DEC OR VICE VERSA

    var timeInfo = new [2];
    var moment;

    // get moment for 2021-12-01 16:00:00 UTC
    moment = Gregorian.moment({
        :year   => 2021,
        :month  => 12,
        :day    => 1,
        :hour   => 16,
        :minutes => 0,
        :seconds => 0
    });
    timeInfo[0] = Gregorian.info(moment, Time.FORMAT_MEDIUM);

    // get moment for 2021-07-01 16:00:00 UTC
    moment = Gregorian.moment({
        :year   => 2021,
        :month  => 7,
        :day    => 1,
        :hour   => 16,
        :minutes => 0,
        :seconds => 0
    });
    timeInfo[1] = Gregorian.info(moment, Time.FORMAT_MEDIUM);

    return timeInfo[0].hour == timeInfo[1].hour;
}

TL;DR Install placeholder app in sim only and use System.isAppInstalled() to test for placeholder app/sim (CIQ 3.2 and higher)

--

Another idea, for CIQ 3.2 and higher:

1a) Create a placeholder app with known UUID whose only purpose is to be "installed" in the simulator

1b) To avoid accidentally sideloading this app, build it for a device type you don't own. (Newer devices will apparently discard sideloads that were built for a different device)

1c) Install the app in the sim by building and running it once

1d) OPTIONAL: if you are concerned about the tiny possibility of the placeholder app's UUID being accidentally duplicated by a real app, upload the placeholder app to the store as a beta. It will available to you and nobody else, so all you have to do is be sure to never install on it a real device (which is why it's helpful to specify a device in manifest.xml that you don't own and never will, like oregon_7xx or something). Once your placeholder app is in the store, it should be impossible for anyone else to use that UUID for a new store app. Of course, that doesn't prevent anyone from sideloading an app that coincidentally has the same UUID

2) To test for the simulator, use System.isAppInstalled() and pass in the placeholder app's UUID

Obviously this will be annoying if you ever delete all apps in the simulator, but it's a lot better than the other idea.

Issues:

- it won't work for really old devices (with CIQ < 3.2). Those devices are at least 8 years old at this point though.

- no way to *guarantee* that some random user didn't sideload an app that somehow has the same UUID as your placeholder app