chore: drop legacy comments and the last unused dead_code allows

* io/xref.rs — replaced the 7-line "BlockRecord::is_loaded turned out to
  be unreliable" rationale with a one-liner. The reverted experiment
  lives in git history (e88a946, 0996b75).
* refedit::RefEditSession.insert_handle — set at construction but never
  read; the "kept for future REFCLOSE DISCARD" TODO can come back via
  git history when that feature lands.
* scene::mod — dropped the placeholder MeshModel re-export; nothing
  outside `mesh_model` consumes the bare name and MeshLodSet is the
  pipeline-facing type now.

Co-Authored-By: Claude Opus 4.7 <[email protected]>

src/io/xref.rs

@@ -37,15 +37,7 @@ pub struct XrefInfo {
 ///
 /// Returns a list of [`XrefInfo`] describing each xref block found.
 pub fn resolve_xrefs(doc: &mut CadDocument, base_dir: &Path) -> Vec<XrefInfo> {
-    // Collect xref blocks: (name, raw_path, block_record_handle)
-    //
-    // Note: `BlockRecord::is_loaded` (the DWG R2000+ "loaded" bit) turned out
-    // to be unreliable as a signal of AutoCAD's Unload state — different
-    // files with xrefs the manager calls "Loaded" carry both `Some(true)`
-    // and `Some(false)` on disk. Until we identify the actual on-disk
-    // representation of XREF→Unload (likely in an XREF dependency record
-    // or ACAD_XREF_PANEL_DATA xrecord), we auto-resolve every xref and
-    // rely on frustum culling to keep the GPU bounded.
+    // Auto-resolve every xref — frustum + LOD culling keep GPU cost bounded.
     let xref_entries: Vec<(String, String, Handle)> = doc
         .block_records
         .iter()