Bugzilla

Thanks, based on recent comments I think I now understand how this can be regression.

A site can have code like this:

```
if ("visualViewport" in window) {
  // register a listener for the visualViewport's resize event
} else {
  // do something else, e.g. register a listener for the regular resize event
}
```

The expected behaviour would then be that, if an iframe is resized, then either the regular resize listener or the visualViewport resize listener fires.

With visualViewport disabled, the second branch was taken so the regular resize listener is registered and fires. With visualViewport enabled, the first branch is taken but, due to this bug, no event is fired.

That would indeed be a regression. I was just confused because the code in comment 0 didn't show code like this, it only talked about visualViewport.

Thanks, based on recent comments I think I now understand how this can be regression.

A site can have code like this:

```
if ("visualViewport" in window) {
  // register a listener for the visualViewport's resize event
} else {
  // register a listener for the regular resize event
}
```

The expected behaviour would then be that, if an iframe is resized, then either the regular resize listener or the visualViewport resize listener fires.

With visualViewport disabled, the second branch was taken so the regular resize listener is registered and fires. With visualViewport enabled, the first branch is taken but, due to this bug, no event is fired.

That would indeed be a regression. I was just confused because the code in comment 0 didn't show code like this, it only talked about visualViewport.

Thanks, based on recent comments I think I now understand how this can be regression.

A site can have code like this:

```
if ("visualViewport" in window) {
  // register a listener for the visualViewport's resize event
} else {
  // register a listener for the regular resize event
}
```

The expected behaviour would then be that, if an iframe is resized, then either the regular resize listener or the visualViewport resize listener fires.

With visualViewport disabled, the second branch was taken so the regular resize listener is registered and fires. With visualViewport enabled, the first branch is taken but, due to this bug, no event is fired.

That would indeed be a regression. I was just confused because the page linked in comment 0 didn't contain code like this, it only used visualViewport.