Position: getScrollInfo() swapped width and height

[CR]

I wanted to use the scrollbar detection/calculation of jquery-ui-position and wondered why it switched the results.

In jquery.ui.position.js line 66 says:

return {
  width: hasOverflowX ? $.position.scrollbarWidth() : 0,
  height: hasOverflowY ? $.position.scrollbarWidth() : 0
};

If I have an overflowing content in horizontal (x) direction, I get a scrollbar on the bottom of the element, that means the height value has been modified. So in my opinion it should return the scrollbarwidth as the height attribute and not as the width.

[Scott González]

It looks to me like the code does exactly what you're saying. Also, it sounds like you're not even reporting a bug against position, but asking a question about how to use some of position's code outside of position. Can you please provide a reduced test case showing the problem you're having?

[CR]

Think of a simple page which has scrollbars on the right. That means, the content's height is too large to fit in. Code would correctly detect hasOverflowY=true and return somthing like {width:0,height:20}.

Now for example I would want to calculate the coordinates of an element to position it. And for that I would need the width of that (possible) scrollbar on the right. So I would expect to find this value in "width" attribute of the result instead of the "height".

I don't know If that affects other calculations in position or the "mind-switching" of the values is going through the rest of the code. For me this just doesn't seem to be right, maybe my way of thinking just differs from yours...

[Scott González]

It sounds like you don't have a bug to report. If you have a test case that is failing, please let us know.

[mikesherov]

I'd like to take a shot at proving the bug here. jsfiddle coming at some point some.

[Scott González]

Sorry, reading this (and the code) again, it does sound like the logic might be reversed.

[Jörn Zaefferer]

Switching around hasOverflowX and hasOverflowY doesn't seem to make a difference in practice, at least for the position utility. Does someone have a case where it actually produces the wrong result?

[mikesherov]

That's what I'm looking into

[Scott González]

Position: Fix orientation check for scrollbar widths. Fixes #8763 - Position: getScrollInfo() swapped width and height.

Changeset: e9c04bfa430eb6b18e7fe1be2f8d162e01181a94