Add free-threaded CPython 3.13t to CI #12550
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
zzzeek
requested changes
Apr 22, 2025
| @@ -1585,7 +1586,7 @@ def predictable_gc(self): | ||
| gc.collect() is called, as well as clean out unreferenced subclasses. | ||
|
|
||
| """ | ||
| return self.cpython | ||
| return self.cpython and not bool(sysconfig.get_config_var("Py_GIL_DISABLED")) |
There was a problem hiding this comment.
i would make a new requirement (that is, a function inside this file) called gil_enabled and use that where needed
test/aaa_profiling/test_memusage.py
Outdated
| @@ -54,6 +56,11 @@ | ||
| from ..orm import _fixtures | ||
|
|
||
|
|
||
| pytestmark = pytest.mark.skipif( |
test/aaa_profiling/test_memusage.py
Outdated
| @@ -1051,6 +1058,12 @@ def test_many_discarded_relationships(self): | ||
| """a use case that really isn't supported, nonetheless we can | ||
| guard against memleaks here so why not""" | ||
|
|
||
| import sysconfig |
There was a problem hiding this comment.
not needed, use gil_enabled req on test_many_discarded_relationships
e.g. for this one and the rest:
@testing.requires.gil_enabled
def test_many_discarded_relationships()
test/base/test_events.py
Outdated
| import sysconfig | ||
| import pytest | ||
|
|
||
| if bool(sysconfig.get_config_var("Py_GIL_DISABLED")): |
| import sysconfig | ||
| import pytest | ||
|
|
||
| if bool(sysconfig.get_config_var("Py_GIL_DISABLED")): |
test/orm/declarative/test_mixin.py
Outdated
| import sysconfig | ||
| import pytest | ||
|
|
||
| if bool(sysconfig.get_config_var("Py_GIL_DISABLED")): |
test/orm/test_events.py
Outdated
| import sysconfig | ||
| import pytest | ||
|
|
||
| if bool(sysconfig.get_config_var("Py_GIL_DISABLED")): |
|
Thanks for the review @zzzeek, I've been running the tests (without any parallelism) using tox, sqlite issues arise when tests are run in parallel |
CaselIT
reviewed
Apr 24, 2025
| @@ -134,6 +135,7 @@ jobs: | ||
| - "3.11" | ||
| - "3.12" | ||
| - "3.13" | ||
| - "3.13t" |
Viicos
reviewed
May 24, 2025
| @property | ||
| def gil_enabled(self): | ||
| """Test should run if the Python distribution is GIL-enabled.""" | ||
| return bool(sysconfig.get_config_var("Py_GIL_DISABLED")) |
There was a problem hiding this comment.
I believe the canonical way to check for this is with sys._is_gil_enabled().
sqla-tester
reviewed
May 24, 2025
There was a problem hiding this comment.
OK, this is sqla-tester setting up my work on behalf of zzzeek to try to get revision 7600297 of this pull request into gerrit so we can run tests and reviews and stuff
|
New Gerrit review created for change 7600297: https://gerrit.sqlalchemy.org/c/sqlalchemy/sqlalchemy/+/5896 |
|
Random q pytest-run-parallel, I'm enabling by using If I just run some of our simplest suites, like the test/sql/test_query suite: "60 tests were not run in parallel because of use of thread-unsafe functionality". It can list out the tests that were parallel or not parallel, but it tells me nothing about what "thread unsafe functionality" is involved. Maybe that's because of sqlite? OK so I tried running them against the psycopg2 driver instead, which is definitely threadsafe: same exact message: "60 tests were not run in parallel because of use of thread-unsafe functionality, " Shouldn't there be a way for the plugin to list out what functionality is thread-unsafe? |
|
per https://github.com/Quansight-Labs/pytest-run-parallel ". If a test was not run in a thread pool because pytest-run-parallel detected use of thread-unsafe functionality, the reason will be printed as well." full output is below, I see no reasons? am I doing something wrong? |
|
@zzzeek thanks for the follow-up, I might need to take a look to the test, maybe there's a function that is declared as thread-unsafe and it's not being reported? |
I think I was running the tests just on main. The pytest parallel plugin ran the tests, made the decision to not parallelize a bunch, and did not express any messaging why it made this decision, in contradiction to what the docs say. So this seems like a bug in pytest parallel. If the tool is going to make implicit decisions, it should log the origin of these decisions |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
See #12513
This PR enables free-threaded CPython 3.13t as a supported target in the CI tests, as outlined in the discussion, this PR only tests sqlalchemy without any kind of parallelism, i.e., using
pytest-run-parallel. Most of the changes are related to test skips when garbage collection timing or memory usage collection takes place, since the free-threaded distribution behaves differently from the GIL-enabled one on these aspects, otherwise, all tests are passing locally.Checklist
This pull request is:
must include a complete example of the issue. one line code fixes without an
issue and demonstration will not be accepted.
Fixes: #<issue number>in the commit messageinclude a complete example of how the feature would look.
Fixes: #<issue number>in the commit messageHave a nice day!