Bugzilla

In bug 1721627 I'm experimenting with making tests fail if we request internal resources that aren't present. The idea is to prevent patches from landing if they mis-spell (or forget to `hg add`, or forget to include in `jar.mn` or similar) subresources like CSS files and images, by failing if we hit chrome: or resource: URLs in tests that are not available (ie where the internal jar/file code cannot find the relevant file).

On android, resource:///modules/AttributionCode.jsm is getting flagged up as a URL that is not present. Some of these are guarded in try...catches, but some are not (though equally

The relevant bits should probably be behind `AppConstants.MOZ_BUILD_APP == "browser"` checks.

Though, to be fair, I also do not see any non-test uses of the ClientEnvironment static `attribution` property, so perhaps removing that altogether would be the easier option?

In bug 1721627 I'm experimenting with making tests fail if we request internal resources that aren't present. The idea is to prevent patches from landing if they mis-spell (or forget to `hg add`, or forget to include in `jar.mn` or similar) subresources like CSS files and images, by failing if we hit chrome: or resource: URLs in tests that are not available (ie where the internal jar/file code cannot find the relevant file).

On android, resource:///modules/AttributionCode.jsm is getting flagged up as a URL that is not present. Some of these are guarded in try...catches, but some are not.
The relevant bits should probably be behind `AppConstants.MOZ_BUILD_APP == "browser"` checks.

Though, to be fair, I also do not see any non-test uses of the ClientEnvironment static `attribution` property, so perhaps removing that altogether would be the easier option?