769848 - window.resizeTo fails is blocked in apps

Status: RESOLVED FIXED

(Firefox Graveyard :: Webapp Runtime, defect, P2)

Description

[Dan Mosedale (:dmosedale, :dmose)]

Tested in an app installed on a Mac Nightly with this UA: 16.0a1 (2012-06-29).

I have game with a canvas in it, and I want to resize the window to exactly the size of the canvas. resizeTo simply does nothing, leaving me with a canvas and a giant ell of whitespace that looks very odd.

This is presumably a consequence of Gecko intentionally blocking it in untrusted content <https://bugzilla.mozilla.org/show_bug.cgi?id=565541#c24>.  I would expect that this restriction wants to be lifted in the more trusted app context.

Comment 1

[Marco Castelluccio [:marco]]

Attachment: allow_window_resize_webapps

The other possible approach is overriding the resizeTo function and calling it again from Chrome code (I'm not sure it would actually work, it's just an idea).

Comment 2

[Marco Castelluccio [:marco]]

Attachment: allow_window_resize_webapps

Forgot to qref...

Comment 3

[Myk Melez [:myk] [@mykmelez]]

Comment on attachment 782053 [details] [diff] [review]
allow_window_resize_webapps


This seems reasonable to me, but I'd like to see what a DOM peer thinks about it.

jst: apps running in the desktop runtime should be able to determine the sizes of their windows, so we're trying to figure out the best way to enable resizeTo in the runtime (without affecting its behavior in desktop Firefox).


(In reply to Marco Castelluccio [:marco] from comment #1)
> The other possible approach is overriding the resizeTo function and calling
> it again from Chrome code (I'm not sure it would actually work, it's just an
> idea).

That might work, but it seems better to make the platform be runtime-aware.

Comment 4

[Johnny Stenback (:jst)]

Comment on attachment 782053 [details] [diff] [review]
allow_window_resize_webapps

In general this seems fine, but this patch as written is not good enough. In particular we must not bypass the drag service code at the end of this changed method. feedback- because of that.

Comment 5

[Marco Castelluccio [:marco]]

Attachment: Patch

Comment 6

[Johnny Stenback (:jst)]

Comment on attachment 787085 [details] [diff] [review]
Patch

   if (mDocShell) {
     bool allow;
     nsresult rv = mDocShell->GetAllowWindowControl(&allow);
+
+    // This preference is useful for the webapp runtime. Webapps should be able
+    // to resize or move their window.
+    if (Preferences::GetBool("dom.always_allow_move_resize_window", false)) {
+      allow = true;
+    }
+

Given that this just ignores what GetAllowWindowControl() says we could just avoid making that call here. I.e. combine the pref check with the mDocShell null check.

r=jst with that changed. I'd happily look over another patch once you have it...

Comment 7

[Marco Castelluccio [:marco]]

Attachment: Patch v2

Comment 8

[Johnny Stenback (:jst)]

Comment on attachment 787134 [details] [diff] [review]
Patch v2

Looks good!

Comment 9

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/fx-team/rev/c279994c083d

Comment 10

[Tim Taubert [:ttaubert] (inactive)]

https://hg.mozilla.org/mozilla-central/rev/c279994c083d