In bug 1494422, we changed the way "initial-scale" affects the ICB size computation.

For example, consider a device width of 800px and a meta viewport tag of:

<meta name="viewport" content="width=device-width; initial-scale=2">

Prior to bug 1494422, this would size the ICB to 800 px, such that at the initial zoom level of 2, the visual viewport would be half the size of the ICB.

With bug 1494422 in place, this sizes the ICB to 400 px (so that it matches the device width at the initial scale), such that at the initial zoom level of 2 the visual viewport and the ICB coincide.

Some APZ tests were relying on the former behaviour to make the visual viewport smaller than the layout viewport. In particular, this is the case for helper_fixed_position_scroll_hittest.html; there may be others as well.

We should audit every usage of initial-scale in an APZ test and make sure it's still doing what we want in light of bug 1494422.

Note that I've gotten rid of some uses of initial-scale, including the one in helper_fixed_position_scroll_hittest.html, in bug 1608506; those removals automatically address this concern for the affected tests.

Hey Botond, Can I work on this issue?

Hey Tanweer, I do not think this is a good beginner bug. It requires an understanding of the mobile viewport sizing model, and a careful look through a number of affected tests to understand their purpose and if their purpose is well served by using initial-scale.

I would generally recommend looking for bugs which have the "Mentor" field set. There is a convenient site for searching these: https://codetribute.mozilla.org/. Note, I think you should feel free to try bugs marked "good first bug" even though it's not your first bug, as we generally have more of these than first-time contributors available to take them.

Okay Thanks for this info. I will look for another bug.