Testcase: data:text/html,<!doctype html><div dir=rtl style="width: 20em" id=foo><span style="border: 2px solid;">Box</span> Some long text that should make sure we line-wrap. If this didn't line-wrap, it should be longer.</div> <input type=range step=0.1 min=10 max=50 value=20 onchange="foo.style.width=this.value+'em';"> You should see that the right border of "Box" is suppressed at first, but it comes back if the box gets wide enough that the text no longer linewraps. The same thing happens if dir=ltr and the text is all rtl, such as in this testcase: data:text/html;charset=utf-8,<!doctype html><div style="width: 20em" id=foo><span style="border: 2px solid;">%D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99</span>%D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99 %D7%A2%D6%B4%D7%91%D6%B0%D7%A8%D6%B4%D7%99</div> <input type=range step=0.1 min=10 max=50 value=20 onchange="foo.style.width=this.value+'em';"> I don't believe we should ever be suppressing the border here. At the very least, no other browser suppresses the border.
Created attachment 460393 [details] Safari 15.5 matches other browsers I am unclear on expected result but Safari 15.5 matches other browsers as can be referred from attached screenshot. Further, I don't see difference on multiple slider level using inline test cases in Comment 01. If this is something, which had web-spec change or this is something now expected behavior and since it is consistent across all browsers. Can this be marked as "RESOLVED CONFIGURATION CHANGED". Thanks!