Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

TST,DOC: Bump scipy_doctest (or remove pin) and fix new failures #28019

Closed
seberg opened this issue Dec 17, 2024 · 6 comments · Fixed by #28020
Closed

TST,DOC: Bump scipy_doctest (or remove pin) and fix new failures #28019

seberg opened this issue Dec 17, 2024 · 6 comments · Fixed by #28020

Comments

@seberg
Copy link
Member

seberg commented Dec 17, 2024

@ev-br ping FYI, since the new scipy-doctest release (less than an hour ago) the refcheck has universal failures.

I suspect, this is all just bad documentation that needs fixing, but not sure yet. In either case, until fixed both CircleCI and the "benchmark" tests which also still run the refcheck are expected to fail.

(Currently, also the linter just started failing...)
@mattip
Copy link
Member

mattip commented Dec 17, 2024

We should pin tools to a specific version, and update in a controlled fasion.

@mattip mattip mentioned this issue Dec 17, 2024
Merged
@mattip
Copy link
Member

mattip commented Dec 17, 2024

I see we do not pin matplotlib, scipy, pytz, pandas. Should we?

@seberg
Copy link
Member Author

seberg commented Dec 17, 2024

I see we do not pin matplotlib, scipy, pytz, pandas. Should we?

Seems fine not to, since I don't think we ever ran into an issue and it seems simple enough to add when we do?
But yes, in this case unless someone wants to sprint on fixing everything, it probably makes sense to add a lower pin and remove after that is OK.

@ev-br
Copy link
Contributor

ev-br commented Dec 17, 2024

Here's failing job: https://app.circleci.com/pipelines/github/numpy/numpy/30210/workflows/1d8f4112-60a7-42f9-ae82-3ff7ed57b6b5/jobs/44299

Two kinds of issues:

The issues at .strides, .shape and .fields (duplicated many times over across dtype classes) are certainly bad/non-ideal examples.

@seberg seberg changed the title TST,DOC: Universal doctest failures (circleCI and "benchmarking") due to scipy_doctest release TST,DOC: Bump scipy_doctest (or remove pin) and fix new failures Dec 17, 2024
@seberg seberg reopened this Dec 17, 2024
@seberg
Copy link
Member Author

seberg commented Dec 17, 2024

Reopening, we really should fix the new failures (and bump the version) soon!

ev-br added a commit to ev-br/scipy that referenced this issue Dec 17, 2024
@ev-br
TST: io.matlab: temp skip failures in io.matlab.Matlab****
9aee37f 
Matlab{Object,Function} inherit from np.ndarray; examples
from {ndarray.strides,shape,fields} are broken. Should
probably keep skipping in scipy until
numpy/numpy#28019 is fixed.
@ev-br ev-br mentioned this issue Dec 17, 2024
Merged
@charris charris mentioned this issue Dec 17, 2024
Merged
@mattip mattip mentioned this issue Dec 17, 2024
Merged
@seberg
Copy link
Member Author

seberg commented Dec 19, 2024

Should have been closed when merging gh-28023

@seberg seberg closed this as completed Dec 19, 2024
tylerjereddy pushed a commit to tylerjereddy/scipy that referenced this issue Dec 20, 2024
@ev-br @tylerjereddy
TST: io.matlab: temp skip failures in io.matlab.Matlab****
e6b2503 
Matlab{Object,Function} inherit from np.ndarray; examples
from {ndarray.strides,shape,fields} are broken. Should
probably keep skipping in scipy until
numpy/numpy#28019 is fixed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

CI: pin scipy-doctest to 1.5.1 mattip/numpy
3 participants
@seberg @mattip @ev-br