`descendants(of:)` recursed on wrapped XPath nodes and, at every node, built
a fresh array for that node's subtree and concatenated it
(`result.append(contentsOf: descendants(of: child))`). That is an array
allocation and copy per node, roughly O(n x depth) work over the subtree, and
the parse profile showed it dominating XPath evaluation for a `//`-style query
over a wide document.
Recurse on the raw tree node instead, appending into one shared inout
accumulator. The result is the identical node-set in the identical document
order; only the way the array is built changes, from O(n x depth) to O(n).
Measured xpath ~46% faster on the 20k-item generated corpus via a same-state
A/B (count unchanged at 10000). Full suite, the Apache Xalan XSLT gold-output
corpus, and the WASM build stay green.
The following and preceding axes work from attribute and namespace
starts, anchored at the owner element per XPath 1.0 document order:
an attribute follows its element, precedes the element's children,
and has no descendants of its own. Xalan baseline 182 -> 174
(axes 16 -> 8).
Rebuild the XPath engine on the parent-aware tree. Add an XPath node
abstraction (tree, attribute, namespace) with document order, and navigate
all thirteen axes: child, descendant, parent, ancestor, following-sibling,
preceding-sibling, following, preceding, attribute, namespace, self,
descendant-or-self, ancestor-or-self. Support the ./../@/// abbreviations
and processing-instruction() node tests. Reverse axes number nearest-first
for positional predicates; results return de-duplicated in document order.
The attribute axis excludes xmlns declarations; the namespace axis surfaces
in-scope namespace nodes.
Closes#40