Loosen tolerances on some tests on i386
authorDebian Science Maintainers <debian-science-maintainers@lists.alioth.debian.org>
Fri, 3 Jul 2026 21:13:01 +0000 (22:13 +0100)
committerRebecca N. Palmer <rebecca_palmer@zoho.com>
Fri, 3 Jul 2026 21:13:01 +0000 (22:13 +0100)
commita31e4b563ea657491eb14551effa4afa6fb1766f
tree1c42638ae742b138f5d9f11cb8e4833709f63747
parenta290020e592226a7f0ec289750d0d0739b679246
Loosen tolerances on some tests on i386

and xfail a test that fails for not producing ConvergenceWarning.

Failure logs:
test_smoothing
https://buildd.debian.org/status/fetch.php?pkg=statsmodels&arch=i386&ver=0.9.0-3&stamp=1567157609&raw=0
test_multivariate_switch_univariate
https://buildd.debian.org/status/fetch.php?pkg=statsmodels&arch=i386&ver=0.12.0-1&stamp=1599693472&raw=0
TestMixedLM.test_compare_numdiff
salsa-ci for 0.14.2-1

As it only fails on i386 and isn't far wrong,
I suspect different rounding due to x87 excess precision

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Bug-Debian: partly https://bugs.debian.org/938949
Forwarded: no

Gbp-Pq: Name i386_loosen_test_tolerances.patch
statsmodels/discrete/tests/test_truncated_model.py
statsmodels/iolib/tests/test_summary2.py
statsmodels/regression/tests/test_lme.py
statsmodels/tsa/holtwinters/tests/test_holtwinters.py
statsmodels/tsa/statespace/tests/test_exact_diffuse_filtering.py
statsmodels/tsa/statespace/tests/test_multivariate_switch_univariate.py
statsmodels/tsa/statespace/tests/test_smoothing.py