Add CI support for Python 3.14#15224
Conversation
|
One or more of the following people are relevant to this code:
|
|
Ha, we need to bump |
Pull Request Test Coverage Report for Build 23020650902Details
💛 - Coveralls |
mtreinish
left a comment
There was a problem hiding this comment.
This LGTM, I'll update the branch protection rules and then enqueue this for merge
|
As discussed offline let's try skip the failing test until python/cpython#140559 & co are resolved, rather than not testing on latest Python. We'll target this for after rc1 though. |
098f46a to
edaa711
Compare
raynelfss
left a comment
There was a problem hiding this comment.
A couple of quick questions before I hit merge. This looks straightforward to me.
| "environment_type": "virtualenv", | ||
| "show_commit_url": "https://github.com/Qiskit/qiskit/commit/", | ||
| "pythons": ["3.10", "3.11", "3.12", "3.13"], | ||
| "show_commit_url": "http://github.com/Qiskit/qiskit/commit/", |
There was a problem hiding this comment.
Why did we go from https to http?
There was a problem hiding this comment.
Thanks, this is probably a bad merge-conflict resolution from a rebase over #15629
| [tool.black] | ||
| line-length = 100 | ||
| target-version = ['py310', 'py311'] | ||
| target-version = ['py310', 'py311', 'py312', 'py313'] |
There was a problem hiding this comment.
Should we include 'py314' here or is black not compatible yet?
There was a problem hiding this comment.
The version of black we use isn't compatible - I tried but had to revert.
There was a problem hiding this comment.
Well, it wasn't at the time, at least. Then I merged #15225, and now it's been a few months, so we might actually be able to update it. I'll leave it for this PR, though, since it'd involve the separate (mostly unrelated) change of bumping the black dependency.
I previously approved the PR. However, I noticed we didn't update ctest.yml and docs-deploy.yml to 3.14.
We already build wheels that are compatible with Python 3.14, so this moves our "latest supported" version in CI to reflect that. There is one visualisation test that was asserting an exact string match of a particular Numpy array that has changed the sign of some zero. This is irrelevant to the actual test (which seems to be testing internals of a method anyway), so we can just make it match regardless of signedness.
edaa711 to
6768030
Compare
mtreinish
left a comment
There was a problem hiding this comment.
LGTM, it's definitely simpler with the new structure
|
By the way:
As of #15794 |
We already build wheels that are compatible with Python 3.14, so this moves our "latest supported" version in CI to reflect that.
There is one visualisation test that was asserting an exact string match of a particular Numpy array that has changed the sign of some zero. This is irrelevant to the actual test (which seems to be testing internals of a method anyway), so we can just make it match regardless of signedness.
Summary
Details and comments
Fix #15135