Judge a renamed element by the name the author wrote (#445, #450)

[jmanico]

Fixes #445 and #450, which share one cause: writeOpenTag in the policy class judged a renamed element by the name the ElementPolicy emitted, while the lexer, the tag balancer, deferOpenTag and disallowTextIn all go by the name the author wrote.

#445 allowElements(policy, "span") registers span as a text container, but the text gate looked up the adjusted name, so span renamed to div without allowElements("div") came out as <div></div>. The gate now follows the input name. One hold-out: a rename into an element whose content a browser reads literally, such as style, still needs allowTextIn on that name, so a rename cannot get around the bar the builder sets for raw-text elements.

#450 The push onto the policy's open-element stack keyed on the adjusted name, so br renamed to span left an entry no close tag ever popped. The span stayed open until its parent closed and swallowed every sibling, and because the balancer never counts a void input, 3000 <br> came out as 3000 nested spans, past the balancer's cap of 256. The push now keys on the input name: a void input renamed to a non-void output is emitted and closed at once, giving <p>a<span></span>bcd</p><p>e</p> for the issue's example.

The reverse rename, span to br, now gets a stack entry with nothing to close, so its own close tag pops it. Before, the close tag found the nearest open element of the same input name and ended an outer span early; and disallowTextIn("span") now holds for it, as it does for every other renamed element.

Seven regression tests in HtmlPolicyBuilderTest pin each shape, including the nesting case. Both existing rename tests keep passing unchanged. Verified with ./mvnw clean verify on JDK 11, 17, 21 and 25.

🤖 Generated with Claude Code

https://claude.ai/code/session_014Ydng7gBm6Ax5zfwt4vZip