pub fn resolve_toctree(
env: &BuildEnvironment,
docname: &str,
toctree: &Node,
) -> Option<Node>Expand description
Resolve a toctree node into the nested bullet lists Sphinx writes in
its place: _resolve_toctree (adapters/toctree.py:119-220) with the
arguments the write phase passes for a document’s own toctrees
(get_and_resolve_doctree: prune=True, the node’s own maxdepth,
collapse=False, includehidden=False).
None is Sphinx’s “renders to nothing”: a hidden toctree, one with no
entries, or one whose entries all name documents without a title.
The result carries exactly what the search index reads from the resolved
tree — the caption as a title, every entry’s title text inside
reference nodes, pruned to maxdepth/titlesonly/tocdepth like
Sphinx prunes it. Still missing for the HTML writer: the
toctree-l<N>/current classes and iscurrent marks, the relative
refuris, and the circular toctree references detected warning (the
circular entry is dropped as Sphinx drops it, but nothing is logged).
only nodes are kept with their children: Sphinx evaluates them against
the build tags, which this crate does not evaluate anywhere yet.