matthew-brett / nipy

87%

placeholder: 83%

LAST BUILD ON BRANCH fix-numpy-1.13-breakage

branch: fix-numpy-1.13-breakage

CHANGE BRANCH

x

Reset

• fix-numpy-1.13-breakage
• add-coveralls
• add-dct-ii
• ants-wrappers
• better-docs-for-testing
• contrast-futz
• convolve-kernels
• convolve-review
• debugging-matrix-rank
• depend-sympy-0.7.0
• docker-build
• examples-refactor
• fix-32-bit-precision-error
• fix-design-errors
• fix-ds105-example
• fix-equals-none
• fix-examples
• fix-formula-errors
• fix-interpolation-error
• fix-matrix-rank-test
• fix-mroi
• fix-select-roi-again
• fix-update-doc-build
• fix-viz-error
• gmm-plot-2d-fix
• hackaround-for-intp-bug
• hrfs-cleanup
• intermittent-imm
• matplotlib-3.4-fix
• more-py3-fixes
• move-data-package-example
• nibabel-compatibility
• no-mpl-registration
• np-12-compat
• np-pre-check-fix
• numpy-1.10-fix
• numpy-114-fixes
• openfmri-helpers
• prefer-numpy-pinv
• prepare-0.4.1
• py3-examples
• refactor-image-time-diff
• refactor-mroi
• refactor-reggie-affine
• refactor-setup.py
• reggy-affine-again
• relax-doc-requirements
• release-preparations
• release-preps
• remove-fmristat-delay
• remove-np-object
• reverting-lapack-refactor
• rm-nipy-devel
• scipy-0.17-error
• setup-helper-fix
• setuptools-always
• shebangs-again
• stack-designs
• sympy-1.0-compat
• test-pre-wheels
• testing-interpolate
• tmp
• to-entry-points
• to-pytest
• to-versioneer
• travis-scipy-fix
• tweak-travis-testing
• update-min-requirements
• update-readme
• use-scriptrunner
• win-27-ci
• windows-27-ci-fixes

Build # 541

Build Type

push

travis-ci

Committed by matthew-brett

Commit Message

BF: new numpy protects polynomial coefficients

New numpy returns a copy of the polynomial coefficients from poly.c,
rather than the actual coefficent array, so when we were modifying the
polynomial coefficients, the polynomial object wasn't seeing our
changes, leading to errors in the RFT module - see e.g.

https://nipy.bic.berkeley.edu/builders/nipy-py2.7-osx-10.10/builds/52/steps/shell_9/logs/stdio

Numpy issue raised on numpy-discussion mailing list.

Run Details

24472 of 28240 relevant lines covered (86.66%)

0.87 hits per line