Each thread specializes a thread-local copy of the bytecode, created on the first RESUME, in free-threaded builds. All copies of the bytecode for a code object are stored in the co_tlbc array on the code object. Threads reserve a globally unique index identifying its copy of the bytecode in all co_tlbc arrays at thread creation and release the index at thread destruction. The first entry in every co_tlbc array always points to the "main" copy of the bytecode that is stored at the end of the code object. This ensures that no bytecode is copied for programs that do not use threads.
Thread-local bytecode can be disabled at runtime by providing either -X tlbc=0 or PYTHON_TLBC=0. Disabling thread-local bytecode also disables specialization.
Concurrent modifications to the bytecode made by the specializing interpreter and instrumentation use atomics, with specialization taking care not to overwrite an instruction that was instrumented concurrently.
* Remove `@suppress_immortalization` decorator
* Make suppression flag per-thread instead of per-interpreter
* Suppress immortalization in `eval()` to avoid refleaks in three tests
(test_datetime.test_roundtrip, test_logging.test_config8_ok, and
test_random.test_after_fork).
* frozenset() is constant, but not a singleton. When run multiple times,
the test could fail due to constant interning.
Run them with different locales and different date and time.
Add the @run_with_locales() decorator to run the test with multiple
locales.
Improve the run_with_locale() context manager/decorator -- it now
catches only expected exceptions and reports the test as skipped if no
appropriate locale is available.
Fix the incorrect use of `os.open()` result as a context manager,
while it is actually a numeric file descriptor.
I have missed the problem, because in the original version the
`os.open()` call would always fail, and I failed to test the final
version in all possible scenarios properly.
Change the default multiprocessing start method away from fork to forkserver or spawn on the remaining platforms where it was fork. See the issue for context. This makes the default far more thread safe (other than for people spawning threads at import time... - don't do that!).
Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com>
Add a helper function that checks whether the test suite is running
inside a systemd-nspawn container, and skip the few tests failing
with `--suppress-sync=true` in that case. The tests are failing because
`--suppress-sync=true` stubs out `fsync()`, `fdatasync()` and `msync()`
calls, and therefore they always return success without checking for
invalid arguments.
Call `os.open(__file__, os.O_RDONLY | os.O_SYNC)` and check the errno to
detect whether `--suppress-sync=true` is actually used, and skip
the tests only in that scenario.
There were a still a number of gaps in the tests, including not looking
at all the builtin types and not checking wrappers in subinterpreters
that weren't in the main interpreter. This fixes all that.
I considered incorporating the names of the PyTypeObject fields
(a la gh-122866), but figured doing so doesn't add much value.
The tests were only checking cases where the slot wrapper was present in the initial case. They were missing when the slot wrapper was added in the additional initializations. This fixes that.
* Move get_signal_name() from test.libregrtest to test.support.
* Use get_signal_name() in support.script_helper.
* support.script_helper now decodes stdout and stderr from UTF-8,
instead of ASCII, if a command failed.
The tests were accidentally disabled by 2da0dc0, which didn't handle classes correctly.
I considered updating no_rerun() to support classes, but the way test_datetime.py works would have made things fairly messy. Plus, it looks like the refleaks we had encountered before have been resolved.
The free-threaded build currently immortalizes objects that use deferred
reference counting (see gh-117783). This typically happens once the
first non-main thread is created, but the behavior can be suppressed for
tests, in subinterpreters, or during a compile() call.
This fixes a race condition involving the tracking of whether the
behavior is suppressed.
We already intern and immortalize most string constants. In the
free-threaded build, other constants can be a source of reference count
contention because they are shared by all threads running the same code
objects.
The code for Tier 2 is now only compiled when configured
with `--enable-experimental-jit[=yes|interpreter]`.
We drop support for `PYTHON_UOPS` and -`Xuops`,
but you can disable the interpreter or JIT
at runtime by setting `PYTHON_JIT=0`.
You can also build it without enabling it by default
using `--enable-experimental-jit=yes-off`;
enable with `PYTHON_JIT=1`.
On Windows, the `build.bat` script supports
`--experimental-jit`, `--experimental-jit-off`,
`--experimental-interpreter`.
In the C code, `_Py_JIT` is defined as before
when the JIT is enabled; the new variable
`_Py_TIER2` is defined when the JIT *or* the
interpreter is enabled. It is actually a bitmask:
1: JIT; 2: default-off; 4: interpreter.
Deferred reference counting is not fully implemented yet. As a temporary
measure, we immortalize objects that would use deferred reference
counting to avoid multi-threaded scaling bottlenecks.
This is only performed in the free-threaded build once the first
non-main thread is started. Additionally, some tests, including refleak
tests, suppress this behavior.
The free-threaded build does not currently support the combination of
single-phase init modules and non-isolated subinterpreters. Ensure that
`check_multi_interp_extensions` is always `True` for subinterpreters in
the free-threaded build so that importing these modules raises an
`ImportError`.
The test suite fetches the C recursion limit from the _testcapi
extension module. Test extension modules can be disabled using the
--disable-test-modules configure option.
These helpers make it easier to customize and inspect the config used to initialize interpreters. This is especially valuable in our tests. I found inspiration from the PyConfig API for the PyInterpreterConfig dict conversion stuff. As part of this PR I've also added a bunch of tests.
On Windows in release mode, the test_cext and test_cppext can now
build C and C++ extensions.
* test_cext now also builds the C extension without options.
* test_cppext now also builds the C++ extension without options.
* Add C++14 test to test_cppext; C++11 is not supported by MSVC.
* Make setup_venv_with_pip_setuptools_wheel() quiet when
support.verbose is false. Only show stdout and stderr on failure.
The free-threaded GC only does full collections, so it uses a threshold that
is a maximum of a fixed value (default 2000) and proportional to the number of
live objects. If there were many live objects after the previous collection,
then the threshold may be larger than 10,000 causing
`test_indirect_calls_with_gc_disabled` to fail.
This manually sets the threshold to `(1000, 0, 0)` for the test. The `0`
disables the proportional scaling.
Add a new C extension "_testlimitedcapi" which is only built with the
limited C API.
Move heaptype_relative.c and vectorcall_limited.c from
Modules/_testcapi/ to Modules/_testlimitedcapi/.
* configure: add _testlimitedcapi test extension.
* Update generate_stdlib_module_names.py.
* Update make check-c-globals.
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
A few of our tests measure the time of CPU-bound operation, mainly
to avoid quadratic or worse behaviour.
Add a helper to ignore GC and time spent in other processes.
Add test annotations required to run the test suite on iOS (PEP 730).
The majority of the change involve annotating tests that use subprocess,
but are skipped on Emscripten/WASI for other reasons, and including
iOS/tvOS/watchOS under the same umbrella as macOS/darwin checks.
`is_apple` and `is_apple_mobile` test helpers have been added to
identify *any* Apple platform, and "any Apple platform except macOS",
respectively.
mpage