1633883 - `TelemetryTestUtils.assertScalar` will probably do the wrong thing if `value` is `0`

Status: RESOLVED FIXED

(Toolkit :: Telemetry, defect, P3)

Description

[Chris H-C :chutten]

TelemetryTestUtils.assertScalar checks if the value passed in is > 0 before asserting that the value is equal. This will not ensure that, e.g., a bool scalar is false.

Maybe we should fix that.

assertKeyedScalar doesn't do it this way.

Comment 1

[Robin Steuber (they/them) [:bytesized]]

This almost caused me trouble. I was planning to encode errors as the value 0, and I was writing a test to verify this functionality. Because of this unusual behavior, I will probably use a different value to convey errors.

Comment 2

[Chris H-C :chutten]

To instead work around the error, please feel free to use the following workaround:

Assert.ok(scalarName in scalars && scalars[scalarName] === value, msg)

Comment 3

[Chris H-C :chutten]

Huh, and the docstring suggests that the function only works for uint scalars:

@param {Number} value The expected value for the provided scalar.

Curious.

Comment 4

[Ed Lee :Mardak]

Attachment: Bug 1633883 - TelemetryTestUtils.assertScalar assumes Number (so behaves odd for boolean, strings) r=chutten!

Allow both 0 and undefined to have TelemetryTestUtils.assertScalar check for not reported scalar.

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/1fc9f15ae2af
TelemetryTestUtils.assertScalar assumes Number (so behaves odd for boolean, strings) r=chutten

Comment 7

[Noemi Erli[:noemi_erli]]

https://hg.mozilla.org/mozilla-central/rev/1fc9f15ae2af