Bugzilla

Thanks Botond for weighing in on this! I'd like to try to have RDM's touch simulation imitate the experience of a real mobile phone using GeckoView as much as possible.  I briefly looked over your second approach at https://bugzilla.mozilla.org/show_bug.cgi?id=1637135#c20:

> 2. Where a page does not explicitly specify an ICB width via a meta viewport tag, change our fallback ICB width from `980px` to `max(980px, screen width)`.
>     * This would prevent the problematic frame trees from arising in the first place, and fix this bug.
>     * There is [some evidence](https://bugzilla.mozilla.org/show_bug.cgi?id=1564253#c6) that Chrome does something similar, namely they use a fallback ICB width of 1280px on some form factors.
>     * While I think this change is unlikely to have adverse effects, it's still a behaviour change with the possibility of unintended consequences, making this a riskier approach than (1).
>     * (Pages that explicitly set an ICB width narrower than the screen width, e.g. that put `width=980` in the meta viewport tag on a 1280px width device, would continue to be broken. However, I expect such pages are very rare.)
> 

To your third point, I wonder if we could just add another check for if RDM is active to avoid introducing new bugs if we decide to change the fallback ICB width to `max(980px, screen width)`. Would that be enough?

Thanks Botond for weighing in on this! If this issue is closely related to what Bug 1637135 details, then I'd like to try to have RDM's touch simulation imitate the experience of a real mobile phone using GeckoView as much as possible.  I briefly looked over your second approach at https://bugzilla.mozilla.org/show_bug.cgi?id=1637135#c20:

> 2. Where a page does not explicitly specify an ICB width via a meta viewport tag, change our fallback ICB width from `980px` to `max(980px, screen width)`.
>     * This would prevent the problematic frame trees from arising in the first place, and fix this bug.
>     * There is [some evidence](https://bugzilla.mozilla.org/show_bug.cgi?id=1564253#c6) that Chrome does something similar, namely they use a fallback ICB width of 1280px on some form factors.
>     * While I think this change is unlikely to have adverse effects, it's still a behaviour change with the possibility of unintended consequences, making this a riskier approach than (1).
>     * (Pages that explicitly set an ICB width narrower than the screen width, e.g. that put `width=980` in the meta viewport tag on a 1280px width device, would continue to be broken. However, I expect such pages are very rare.)
> 

To your third point, I wonder if we could just add another check for if RDM is active to avoid introducing new bugs if we decide to change the fallback ICB width to `max(980px, screen width)`. Would that be enough?