remove() and replaceAll() still match the nodes in hand, unlike find() #126

Once #73 lands, find() searches descendants only, as jQuery's .find() does. remove($selector) and replaceAll($selector) still go through the legacy engine, QueryPath\CSS\QueryPathEventHandler, which counts the nodes in the match set as candidates for their own selector. So the two engines disagree about which nodes a selector reaches:

$xml = '<?xml version="1.0"?><root><p><i>a</i><b>x</b></p><p><i>b</i></p><i>c</i></root>';

qp($xml, 'p')->remove('p')->count();   // 2: both <p> themselves
qp($xml, 'p')->detach('p')->count();   // 0: no <p> below a <p>

qp($xml, 'p')->remove('*')->count();   // 5: the two <p> plus their three children
qp($xml, 'p')->detach('*')->count();   // 3: the three children

detach($selector) was fixed in #108 (issue #76) to detach "the matches of $selector, as remove($selector) does". With #73 that only holds for selectors that cannot match the nodes in hand.

Cause

Every matcher in QueryPathEventHandler (element(), elementID(), elementClass(), anyElement(), elementNS(), anyElementInNS(), attribute(), …) iterates candidateList(), which includes each node in the match set as well as its descendants. anyElement() even says // Add self.

Options

• Make candidateList() descendants-only, with the same document-element exception DOMTraverser::isDocumentElement() uses in Make find() search descendants only, as jQuery does #73, so remove()/replaceAll() agree with find().
• Or move remove()/replaceAll() onto DOMTraverser, removing the last callers of the legacy engine. Related: remove() and replaceAll() throw "Unknown Pseudo-Class: matches" for :matches() #123 and remove('li:not(:first)') and other positional :not() arguments are evaluated per element in the legacy engine #124 are other places where the legacy engine disagrees with find().

DOMQueryTest::testDetachWithSelectorAgreesWithRemove currently skips the self-matching selectors. Restore them once this is fixed.

🤖 Generated with Claude Code