<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8"/>
<title>HTML Test: BDI: paragraph-level container</title>
<link rel="match" href="bdi-neutral-separate-ref.html"/>
<link rel="author" title="Aharon Lanin" href="mailto:aharon@google.com"/>
<link rel="author" title="HTML5 bidi test WG" href="mailto:html5bidi@googlegroups.com"/>
<link rel="help" href="https://html.spec.whatwg.org/multipage/#the-bdi-element"/>
<meta name="assert" content="
'For the purposes of applying the bidirectional algorithm to the paragraph-level
container that a bdi element finds itself within, the bdi element must be treated
like a U+FFFC OBJECT REPLACEMENT CHARACTER.'
Thus, under no circumstances should any part of the content outside a BDI be visually
reordered inside the BDI's content."/>
<style>
body{
font-size:2em;
}
.test, .ref {
border: medium solid gray;
width: 500px;
margin: 20px;
}
.comments { display: none; }
</style>
</head>
<body>
<div class="instructions"><p>Test passes if the two boxes below look exactly the same.</p></div>
<div class="comments">
Key to entities used below:
א ... ו - The first six Hebrew letters (strongly RTL).
‭ - The LRO (left-to-right override) formatting character.
‬ - The PDF (pop directional formatting) formatting character; closes LRO.
If the BDIs in the test's first DIV were just SPANs, the א would appear between the 1
and the b, and the ו between the d and the ה.
</div>
<div class="test">
<div dir="ltr">א <bdi>[1 b]</bdi> c <bdi>[d ה]</bdi> ו...</div>
<div dir="rtl">a <bdi>[1 ב]</bdi> ג <bdi>[ד e]</bdi> f...</div>
</div>
<div class="ref">
<div dir="ltr">‭א [1 b] c [d ה] ו...‬</div>
<div dir="rtl">‭...f [e ד] ג [ב 1] a‬</div>
</div>
</body>
</html>