Skip to content

TST: use double precision for CDF endpoint checks - #17

Merged
mj-will merged 1 commit into
glasflowfrom
tst/cdf-endpoint-precision
Nov 17, 2026
Merged

mj-will merged 1 commit into
glasflowfrom
tst/cdf-endpoint-precision

Conversation

@mj-will

@mj-will mj-will commented Nov 17, 2026

Copy link
Copy Markdown

The TestPiecewiseCDF.test_ones_to_ones check intermittently fails for PiecewiseCubicCDF because float32 evaluation at the upper endpoint can exceed the existing 1e-5 tolerance. This caused the Python 3.12 Ubuntu failure in uofgravity/glasflow#71.

Evaluate this endpoint check in float64, following the approach used for the round-trip checks in #15. Keep the existing tolerance and production implementation unchanged.

@mj-will
mj-will merged commit fe2b584 into glasflow Nov 17, 2026
12 checks passed
@mj-will
mj-will deleted the tst/cdf-endpoint-precision branch November 17, 2026 09:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant