schemagen: reject self-referential array $ref cycles instead of crashing
CI / build (-DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++, clang-arm64, ubuntu-latest-arm64, true) (pull_request) Successful in 55s
CI / build (-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, gcc-arm64, ubuntu-latest-arm64, false) (pull_request) Successful in 51s
CI / pre-commit (pull_request) Successful in 52s
CI / build (-DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++, clang-amd64, ubuntu-latest-amd64, true) (pull_request) Successful in 1m32s
CI / build (-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++, gcc-amd64, ubuntu-latest-amd64, false) (pull_request) Successful in 1m26s

Cyclic array definitions (directly or through a chain of array $defs)
created a self-referential TArr, which then caused infinite recursion in
base_cpp, storage_cpp, and _walk_arrays.

Object-only cycles are already broken with std::unique_ptr, but
array-only cycles have no object edge for break_cycles to cut.

Detect them after building an array's items by following TArr.elem links
and raise a clear GenError so generation fails gracefully rather than
overflowing the Python stack.

Closes #33
This commit is contained in:
2026-06-30 12:29:37 -04:00
parent e22bc039ae
commit ababd3a8fd
2 changed files with 103 additions and 0 deletions
+24
View File
@@ -235,6 +235,24 @@ class Builder:
"""Names generated internally that must not collide with user types."""
return name in ("Root", "RootScalar")
@staticmethod
def _array_reaches(target, ty):
"""Return True if `target` can be reached from `ty` by following
TArr element types. This detects self-referential array cycles that
cannot be expressed as C++ structs."""
seen = set()
stack = [ty]
while stack:
cur = stack.pop()
if cur is target:
return True
if id(cur) in seen:
continue
seen.add(id(cur))
if isinstance(cur, TArr):
stack.append(cur.elem)
return False
def ref_name(self, ref):
if not ref.startswith("#/"):
raise GenError(f"only local $ref supported, got: {ref}")
@@ -322,6 +340,12 @@ class Builder:
)
t.elem = elem
t.elem_nullable = elem_nullable
# Self-referential array cycles (directly or through a chain of
# array definitions) cannot be represented as a C++ value type.
if self._array_reaches(t, elem):
raise GenError(
"recursive array type is not supported: " f"{defname or hint!r}"
)
return result
scalar = {