Migrate jsonpull to unique_ptr ownership

Replaces the shared_ptr-based json_object_ptr with a unique_ptr that
has a stateless custom deleter dispatching on json_object::type before
calling the right subclass destructor. Eliminates per-node atomic
reference-counting and the control-block allocation that shared_ptr
required for every node in the tree.

API now distinguishes owning and borrowing pointers explicitly:
- json_read / json_read_separators / json_hash_get return raw
  json_object * (borrowed from the parser-owned tree).
- json_read_tree / json_disconnect return json_object_ptr (caller
  takes ownership; back-pointers are cleared so the subtree can
  outlive the parser).
- json_free / json_context / json_stringify take raw pointers.
- The parser's container_stack holds raw pointers; jp->root keeps
  unique_ptr ownership of the most recent top-level value.

Internally, take_from_owner moves the unique_ptr out of whichever
parent vector / hash entry / parser root owned it, which both
json_free and json_disconnect rely on.

In the streaming parsers (parse_feature, parse_layers, the
geojson-loop callback), we are careful to free `j` only after we
have processed a complete Feature: json_read returns each token
as the tree is being built up, and freeing an intermediate node
would splice it out of the surrounding hash and corrupt the
in-progress feature.

Benchmark (tl_2022_us_county.json, -z0 --extend-zooms-if-still-dropping,
median of 5 runs on macOS arm64): 8.5s, vs 10.6s with shared_ptr
and 8.7s on the pre-refactor C baseline.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Erica Fischer
2026-05-30 20:44:25 -07:00
co-authored by Cursor
parent 1a7075e3ab
commit 65beb3f0d7
23 changed files with 408 additions and 289 deletions
+28 -15
View File
@@ -174,10 +174,10 @@ TEST_CASE("jsonpull surrogate-pair regression", "[jsonpull][surrogate]") {
TEST_CASE("json_free prunes a subtree from its parent", "[jsonpull][memory]") {
json_pull_ptr jp = json_begin_string("[[1, 2], [3, 4], [5, 6]]");
json_object_ptr outer;
json_object *outer = nullptr;
int arrays_seen = 0;
json_object_ptr j;
json_object *j;
while ((j = json_read(jp)) != nullptr) {
if (j->type != JSON_ARRAY) {
continue;
@@ -190,7 +190,8 @@ TEST_CASE("json_free prunes a subtree from its parent", "[jsonpull][memory]") {
REQUIRE(j->array()[1]->number() == 4);
json_free(j);
} else if (j->parent == nullptr) {
// The completed outer array.
// The completed outer array; the parser still owns it
// via jp->root, so the borrowed pointer stays valid.
outer = j;
break;
}
@@ -215,21 +216,33 @@ TEST_CASE("json_free prunes a subtree from its parent", "[jsonpull][memory]") {
// The companion case to the pruning test above: in a line-delimited
// stream, each feature returned by json_read is a top-level value
// with no parent, but the parser still co-owns it via jp->root.
// with no parent, but the parser still owns it via jp->root.
// json_free must drop that parser reference too, otherwise the
// just-serialized feature stays in memory until the next feature
// starts parsing.
// just-serialized feature would sit in memory until the next feature
// started parsing. Under the unique_ptr ownership model, the only
// owner is jp->root, so verifying that jp->root is empty after the
// json_free call is also a guarantee that the subtree itself has
// been destroyed.
TEST_CASE("json_free releases a top-level value held by the parser", "[jsonpull][memory]") {
std::weak_ptr<json_object> observer;
json_pull_ptr jp = json_begin_string(R"({"a": 1, "b": [2, 3]})");
json_object_ptr j = json_read_tree(jp);
REQUIRE(j != nullptr);
REQUIRE(j->type == JSON_HASH);
REQUIRE(j->parent == nullptr);
observer = j;
json_free(j);
// json_read streams atoms first (1, 2, 3, [2,3], ...); the top-level
// hash is returned by the final `}` token.
json_object *top = nullptr;
json_object *j;
while ((j = json_read(jp)) != nullptr) {
if (j->parent == nullptr) {
top = j;
break;
}
}
REQUIRE(observer.expired());
REQUIRE(top != nullptr);
REQUIRE(top->type == JSON_HASH);
REQUIRE(jp->root.get() == top);
json_free(top);
// top is dangling now; do not dereference.
REQUIRE(jp->root == nullptr);
}