[IFC][line-clamp] imported/w3c/web-platform-tests/css/css-overflow/we…

…bkit-line-clamp-025.html asserts

https://bugs.webkit.org/show_bug.cgi?id=256142

Reviewed by Antti Koivisto.

Adding the "epsilon wiggle room" (see 263428@main) causes assert when we have exact available space values. In such cases we don't find any overlapping display box even though we know the line content is overflowing (i.e. the last (first) display box must be overflowing).

* Source/WebCore/layout/formattingContexts/inline/display/InlineDisplayLineBuilder.cpp:
(WebCore::Layout::truncateOverflowingDisplayBoxes):

Canonical link: https://commits.webkit.org/263541@main

Source/WebCore/layout/formattingContexts/inline/display/InlineDisplayLineBuilder.cpp

@@ -256,8 +256,8 @@ static float truncateOverflowingDisplayBoxes(const InlineDisplay::Line& displayL
             }
             isFirstContentRun = false;
         }
-        ASSERT(truncateRight.has_value());
-        return truncateRight.value_or(0.f);
+        ASSERT(truncateRight.has_value() || right(boxes.last()) == visualRightForContentEnd);
+        return truncateRight.value_or(right(boxes.last()));
     }
 
     auto truncateLeft = std::optional<float> { };
@@ -276,8 +276,8 @@ static float truncateOverflowingDisplayBoxes(const InlineDisplay::Line& displayL
         }
         isFirstContentRun = false;
     }
-    ASSERT(truncateLeft.has_value());
-    return truncateLeft.value_or(ellipsisWidth) - ellipsisWidth;
+    ASSERT(truncateLeft.has_value() || left(boxes.first()) == visualLeftForContentEnd);
+    return truncateLeft.value_or(left(boxes.first())) - ellipsisWidth;
 }
 
 std::optional<FloatRect> InlineDisplayLineBuilder::trailingEllipsisVisualRectAfterTruncation(LineEndingEllipsisPolicy lineEndingEllipsisPolicy, const InlineDisplay::Line& displayLine, InlineDisplay::Boxes& displayBoxes, bool isLastLineWithInlineContent)