RFR: 8218826: TableRowSkinBase: horizontal layout broken if row has padding
Michael Strauß
mstrauss at openjdk.org
Tue Jun 21 23:35:01 UTC 2022
On Tue, 24 May 2022 21:25:23 GMT, Marius Hanl <mhanl at openjdk.org> wrote:
> This PR fixes a problem, where the layout is broken when a `(Tree)TableRow` has padding.
> As also mentioned in the ticket, the `layoutChildren` method in `TableRowSkinBase` is implemented wrong.
>
> The `layoutChildren` method is responsible for layouting all the `(Tree)TableCells`.
> When the row has padding, it is subtracted on every `(Tree)TableCell` - which is wrong.
>
> Instead the `x` and `y` from `layoutChildren` should be used. (E.g. if `x` is 10 (=padding left+right = 10), then only the first cell should start at 10 and the other cells follow as usual)
> Also the `compute...` methods needs to add the padding as well.
>
> **Example:**
> _Row padding left right 0:_
> [Cell1][Cell2][Cell3]
> _Row padding left right 10:_
> [ 10 ][Cell1][Cell2][Cell3][ 10 ] (`compute...` method also needs to account the padding)
> _Same for top bottom._
>
> When a `fixedCellSize` is set, the padding is currently ignored (also after this PR).
> Therefore, `y` in the `layoutChildren` method is set to 0 for `fixedCellSize`.
>
> This may can be discussed in the mailing list (Could be a follow up). To support padding also when a `fixedCellSize` is set, the `compute...` methods needs to also add the padding when a `fixedCellSize` is set (in the `if` clauses) and the `VirtualFlow` needs to add the padding to every row instead of using the `fixedCellSize` directly (probably at the cost of performance).
modules/javafx.controls/src/test/java/test/javafx/scene/control/skin/TreeTableRowSkinTest.java line 38:
> 36: import javafx.scene.control.TreeTableView;
> 37: import javafx.scene.control.cell.TreeItemPropertyValueFactory;
> 38: import org.junit.After;
Since this is a new file, I suggest using the JUnit5 API.
-------------
PR: https://git.openjdk.org/jfx/pull/800
More information about the openjfx-dev
mailing list