Cursor vertical hover position is off by 1

USE CASE: WHAT DO YOU WANT TO DO?

Make precise selections or view correct labels associated with a specific pixel hovered over.

STEPS TO REPRODUCE AN ISSUE (OR TRIGGER A NEW FEATURE)

Open the 6kx6k test data file
Try to hover the cursor over the first and last rows/columns and observer whether the label displayed for those are the actual first and last row/column.

CURRENT BEHAVIOR

The top-left black pixel of the cursor is always 1 pixel above the returned hover position. In fact, you can hover that top cursor pixel 1 pixel above the matrix and the reported hover position is the top row. If you hover the top-left black cursor pixel on the bottom row of pixels, the cursor "leaves" the IMV pane and no labels are drawn. Hovering over the second row from the bottom yields the bottom row label.

Horizontally, hovering over the first column and last column of pixels is accurate.

EXPECTED BEHAVIOR

Hovering the top-left black pixel of the cursor over anything should reveal labels corresponding to that position.

DEVELOPERS ONLY SECTION

SUGGESTED CHANGE (Pseudocode optional)

I have not yet been able to rule out the OS yet, so I suggest first you see if you can reproduce this. I was using control-scroll to zoom the entire screen to see precisely what I was doing. The concerning factor here is whether a user feels like they can comfortably tell whether they are viewing data at the top or bottom or extreme left & right sides.

If a coding change is made, it would have to be basal so that mouseEntered & mouseExited are corrected as well as the label position reported.

FILES AFFECTED (where the changes will be implemented) - developers only

unknown

LEVEL OF EFFORT - developers only

major

COMMENTS

This is obviously a trivial issue, but it's one of those issues that can affect the trust a user has in an application. Another way of handling this issue would be to provide some sort of visual feedback that tells a user whether they are actually hovering over the first/last row/col or whether there are more hidden off the edge.

Comments (4)