Repository navigation
Replies: 1 comment
|
Test only PRs: #1257, #1301, #1393, #2165, #2513 (its author says the test passes), #2532, #2715, #2762, #3246 and #5775 (both are qualname tests), #3701, #4374, #4581, #5025, #5406, #5544, #5601, #5678, and #5855 (its author found the cause: private inheritance). |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
This is AI generated; being on this list does not mean the fix is good or going to be accepted. Just that we want to review it. :)
🤖 AI text below 🤖
This is a triage of the open PRs from the most recent 35, as of 2026-11-12. PRs merged today are not listed. The assessment comes from the diffs, the PR threads and the CI results. Nothing was run locally.
Ready to merge (CI green, fix is correct)
native_enumduplicate member messageis_methodReady after a small fix
def_submoduleleak): Clang-Tidy fails onmodernize-use-auto. Changeobject this_file = reinterpret_steal<object>(...)toauto.~pythonbufterminate): the new test was inserted between theskipif(emscripten)decorator andtest_threading. The skip now applies to the new test, andtest_threadingruns on Pyodide and fails. Move the test above the decorator.set::addthrows): the one failure is a DNS error, so it needs a rerun. It is a behaviour change, becauseadd()now always returnstrueor throws.py::castthrows instead of returning null with an error set): CI is green and it matches the docs. It is a behaviour change, so it needs a maintainer decision.array_casterwith a mutating__len__): CI is green and the out-of-bounds access is real. It is hardening against a hostile object, so it is low priority.Good, but needs a rebase
pybind11_object()raises instead of aborting): conflicts inclass-inl.h.map): conflicting. The case is contrived, so it is low value.gil_safe_call_oncereset on finalize): the logic looks correct. It conflicts intest_interpreter.cppafter fix: leave ref_small_vector source empty after move #6161 merged.Needs a reviewer
(Mine, so I can't review)
args_convertfor kw-only arguments afterpy::args): rebased today, CI is running. The logic looks correct.SHAREDmodules against libpython in FindPython mode): rebased today. The iOS wheel job fails and other jobs are still running.pybind11.pcprefix and CMP0190 default): draft, looks correct.Not ready
get_object_handlereturns an object, and the cost on the GIL path is not measured.std::optional<std::reference_wrapper<Eigen::...>>gives segmentation fault #5637, and the scope question on fix: avoid dangling reference in std::optional<std::reference_wrapper<T>> arguments #6217 is unanswered. fix: avoid dangling reference in optional/variant of reference_wrapper #6232 constructs the inner caster eagerly, which is the regression already flagged on fix: avoid dangling reference in std::optional<std::reference_wrapper<T>> arguments #6217. fix: avoid dangling reference in std::optional<std::reference_wrapper<T>> arguments #6217 has 27 failing jobs.string_view): the full CI never ran, so it needs workflow approval. It also adds a memoryview per call and behaves differently on PyPy.PYBIND11_USE_SYSTEM_HEADERS): 47 failing jobs, conflicting, and it changes a property onpybind11_headersfrom insidepybind11_add_module.*in signatures): the documented handler escapes every*, which breaks emphasis and bullet lists in user docstrings.Merge order
Some of these PRs append to the same test files, so the second one to merge needs a rebase:
test_class.py: fix: raise instead of aborting when instantiating pybind11_object #6244, fix: throw from py::cast instead of returning null with an error set #6245test_stl.py: fix out-of-bounds access in array_caster on mutating __len__ #6220, fix(stl): reject non-iterable classes named like builtin iterables #6243All reactions