forked from mirrors/gecko-dev
e1d6dd14b9
A potentially better approach could be to call FrameNeedsReflow from nsBCTableCellFrame::SetBorderWidth... But it seems that might over-invalidate quite a lot when recalculating borders, since it seems we reset the relevant ones and then accumulate directly on the frame. If that seems like a better approach, happy to give that a try. This should be less risky for performance changes in general tho. Differential Revision: https://phabricator.services.mozilla.com/D216409
24 lines
320 B
HTML
24 lines
320 B
HTML
<!doctype html>
|
|
<meta charset="utf-8">
|
|
<title>CSS test reference</title>
|
|
<style>
|
|
table {
|
|
border-collapse: collapse;
|
|
}
|
|
|
|
td {
|
|
border-bottom: 10px solid;
|
|
}
|
|
</style>
|
|
<table>
|
|
<thead>
|
|
<tr>
|
|
<td>Something</td>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr>
|
|
<td>Body</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|