Bug 240388 – computeIntrinsicLogicalContentHeightUsing doesn't take intrinsicBorderForFieldset() into account

• Status: RESOLVED
• Resolution: FIXED
• Component: CSS

[Oriol Brufau]

Runt this:

  <fieldset style="height: min-content; border: solid cyan">
    <legend>foo</legend>
    <div style="border: solid fuchsia">bar</div>
  </fieldset>

The fieldset is not big enough, the contents overflow.

Or run this:

  <fieldset id="target" style="height: min-content">
    <legend>foo</legend>
  </fieldset>
  <script>
  console.log(getComputedStyle(target).height);
  </script>

It logs -15.59375px, negative height!

That's because RenderBox::computeIntrinsicLogicalContentHeightUsing uses

        if (intrinsicContentHeight && style().boxSizing() == BoxSizing::BorderBox)
            return intrinsicContentHeight.value() + borderAndPaddingLogicalHeight();
        return intrinsicContentHeight;

This logic should be moved into a virtual method, and override it in RenderBlock to take intrinsicBorderForFieldset() into account.
Just like adjustBorderBoxLogicalHeightForBoxSizing and adjustIntrinsicLogicalHeightForBoxSizing.

[Oriol Brufau]

Created attachment 459312 [details]
Patch

[EWS Watchlist]

This patch modifies the imported WPT tests. Please ensure that any changes on the tests (not coming from a WPT import) are exported to WPT. Please see https://trac.webkit.org/wiki/WPTExportProcess

[Darin Adler]

Comment on attachment 459312 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=459312&action=review

> Source/WebCore/rendering/RenderBox.cpp:3269
>          return intrinsicContentHeight;

After the refactoring, this could just be return std::nullopt, and I think it’s clearer that way.

[EWS]

Committed r294275 (250621@main): <https://commits.webkit.org/250621@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 459312 [details].

[Radar WebKit Bug Importer]

<rdar://problem/93378304>