585680 - Allow scrollable views to scroll past the end of content

Status: RESOLVED WONTFIX

(Core :: Layout, defect)

Description

[Mark Finkle (:mfinkle) (use needinfo?)]

We have a have issues with scrolling on mobile devices:
* Most mobile OSes allow for scrolling past the end of the content (web or XUL) and then the content will snap back when the user releases
* Some mobile OSes float virtual keyboards over content, but do not resize the actual application window. In order to scroll any textbox or focusable elements at the bottom of the page into view, we may need to scroll the content past the end of the page.

Both of these uses cases are currently out of our grasp since the scrollable view will cease scrolling when it reaches the end of the content.

Comment 1

[Robert O'Callahan (:roc) (email my personal email if necessary)]

OK.

You want to change nsGfxScrollFrameInner::GetScrolledRectInternal to add some height to the returned rectangle. That basically work. There might be some assertions or assumptions that we break, but a quick look at the code doesn't show me any.

Comment 2

[Mark Finkle (:mfinkle) (use needinfo?)]

I was playing around with this bug and tried this change to nsGfxScrollFrameInner::GetScrolledRectInternal:

+  return nsRect(x1, y1, (x2 - x1) + aScrollPortSize.width, (y2 - y1) + aScrollPortSize.height);
-  return nsRect(x1, y1, x2 - x1, y2 - y1);

It seemed to work exactly the way I wanted. I could scroll content of anything a full viewport width (or height) beyond the content. I will try to use attributes or CSS properties to control the amount to over scroll in the x and y directions.

In Fennec, the browser content, drawn to the canvas was changed. It was like the extra scroll extent was used to create the document bounding rect or something. Since we try to render the entire content area in the canvas tiles, I saw a lot of whitespace to the right and below the real web content.

Comment 3

[Mark Finkle (:mfinkle) (use needinfo?)]

Maybe an "overScrollBy" API would be better. Android has one:

http://developer.android.com/reference/android/view/View.html#overScrollBy%28int,%20int,%20int,%20int,%20int,%20int,%20int,%20int,%20boolean%29

Comment 4

[Brad Lassey [:blassey] (use needinfo?)]

Attachment: patch

Matt, is this what is needed for the front end?

Comment 5

[Matt Brubeck (:mbrubeck)]

Comment on attachment 553799 [details] [diff] [review]
patch

Yes, this looks good for our use cases.

Comment 6

[Josh Matthews [:jdm]]

layout/xul/base/public/nsIScrollBoxObject.idl needs an IID rev.

Comment 7

[Brad Lassey [:blassey] (use needinfo?)]

Attachment: patch with uuid rev'd

Comment 8

[Robert O'Callahan (:roc) (email my personal email if necessary)]

How does this interact with scrollbar rendering?

What happens if you're overscrolled and a reflow occurs in the content? Seems to me the reflow will snap you back into range, which you probably don't want.

Comment 9

[Brad Lassey [:blassey] (use needinfo?)]

Madhava, are we going to want overscroll for iframes? If not, I'll probably just cancel this review. If so, I need to figure out answers to roc's questions.

Comment 10

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

(In reply to Brad Lassey [:blassey] from comment #9)
> Madhava, are we going to want overscroll for iframes? If not, I'll probably
> just cancel this review. If so, I need to figure out answers to roc's
> questions.

I guess for mobile you don't need that anymore since you have moved to native widgets, but for B2G it will be nice to have overscroll - even for scrollable views that are not iframes.

Comment 11

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Would still be desired for scrollable areas inside Web pages, wouldn't it?

Comment 12

[Rob Arnold [:robarnold]]

(In reply to Vivien Nicolas (:vingtetun) from comment #10)
> (In reply to Brad Lassey [:blassey] from comment #9)
> > Madhava, are we going to want overscroll for iframes? If not, I'll probably
> > just cancel this review. If so, I need to figure out answers to roc's
> > questions.
> 
> I guess for mobile you don't need that anymore since you have moved to
> native widgets, but for B2G it will be nice to have overscroll - even for
> scrollable views that are not iframes.

Would there be any compatibility concerns with enabling this for iframes & overflow:scroll regions? Seems like overscrolling might reveal some content that the page author didn't intend.

Comment 13

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

(In reply to Rob Arnold [:robarnold] from comment #12)
> (In reply to Vivien Nicolas (:vingtetun) from comment #10)
> > (In reply to Brad Lassey [:blassey] from comment #9)
> > > Madhava, are we going to want overscroll for iframes? If not, I'll probably
> > > just cancel this review. If so, I need to figure out answers to roc's
> > > questions.
> > 
> > I guess for mobile you don't need that anymore since you have moved to
> > native widgets, but for B2G it will be nice to have overscroll - even for
> > scrollable views that are not iframes.
> 
> Would there be any compatibility concerns with enabling this for iframes &
> overflow:scroll regions? Seems like overscrolling might reveal some content
> that the page author didn't intend.


From what I understand of the patch there is a new scrollable.overScrollBy(x, y) method. If the web author did not use it there will be no overscroll of the iframe/overflow:scroll regions of the page.


(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #11)
> Would still be desired for scrollable areas inside Web pages, wouldn't it?

Sorry I didn't realize this question was for me. B2G probably wants scrollable area inside iframes since application are nested iframes and the web browser will be a kind of iframe inside an other iframe.

Comment 14

[Rob Arnold [:robarnold]]

(In reply to Vivien Nicolas (:vingtetun) from comment #13)
> (In reply to Rob Arnold [:robarnold] from comment #12)
> > (In reply to Vivien Nicolas (:vingtetun) from comment #10)
> > > (In reply to Brad Lassey [:blassey] from comment #9)
> > > > Madhava, are we going to want overscroll for iframes? If not, I'll probably
> > > > just cancel this review. If so, I need to figure out answers to roc's
> > > > questions.
> > > 
> > > I guess for mobile you don't need that anymore since you have moved to
> > > native widgets, but for B2G it will be nice to have overscroll - even for
> > > scrollable views that are not iframes.
> > 
> > Would there be any compatibility concerns with enabling this for iframes &
> > overflow:scroll regions? Seems like overscrolling might reveal some content
> > that the page author didn't intend.
> 
> 
> From what I understand of the patch there is a new
> scrollable.overScrollBy(x, y) method. If the web author did not use it there
> will be no overscroll of the iframe/overflow:scroll regions of the page.

Ah ok. I thought this was about overscrolling in response to mouse wheel events or inertial scrolling.

Comment 15

[Robert O'Callahan (:roc) (email my personal email if necessary)]

(In reply to Rob Arnold [:robarnold] from comment #12)
> Would there be any compatibility concerns with enabling this for iframes &
> overflow:scroll regions? Seems like overscrolling might reveal some content
> that the page author didn't intend.

For LTR content, Web authors already can't prevent users from scrolling as far right or down as the content extends. So the only new issue would be if there's content above or to the left of the origin. Sometimes authors shove things there to hide them, but usually they put it quite far off to make sure it doesn't accidentally intrude on the page. I could be wrong, but personally I wouldn't worry about it.

Comment 16

[Brad Lassey [:blassey] (use needinfo?)]

Comment on attachment 553864 [details] [diff] [review]
patch with uuid rev'd

its not clear that this is needed anymore

Comment 17

[Brad Lassey [:blassey] (use needinfo?)]

don't want this