Two different subjects now share the test number 1029 on development:
Both are present and both run; nothing is broken. What collides is the number,
which is how the suite is navigated and how scripts/test.sh selects batches by
glob. A range like tests/test_102*py now picks up two unrelated subjects under
one identifier, and anyone citing "test 1029" has to say which.
The two PRs were written in parallel against different bases, so neither could
see the other's choice — #550 merged on 2026-08-17 and #594 was branched before
it.
Fix: renumber the Zhong file to test_1030_analytic_zhong2008.py. The faulted
medium keeps 1029 because it merged first. No content change; check
scripts/test.sh globs and the docs under docs/api/ and
docs/developer/subsystems/ for references to the file name before renaming.
Underworld development team with AI support from Claude Code
Two different subjects now share the test number 1029 on
development:tests/test_1029_analytic_faulted_medium.py— Barr & Houseman's faultedmedium, from Barr & Houseman faulted-medium analytic solution in uw.analytic #550
tests/test_1029_analytic_zhong2008.py— the Zhong 2008 propagator-matrixresponse, from Add Zhong 2008 propagator-matrix analytic response #594
Both are present and both run; nothing is broken. What collides is the number,
which is how the suite is navigated and how
scripts/test.shselects batches byglob. A range like
tests/test_102*pynow picks up two unrelated subjects underone identifier, and anyone citing "test 1029" has to say which.
The two PRs were written in parallel against different bases, so neither could
see the other's choice — #550 merged on 2026-08-17 and #594 was branched before
it.
Fix: renumber the Zhong file to
test_1030_analytic_zhong2008.py. The faultedmedium keeps 1029 because it merged first. No content change; check
scripts/test.shglobs and the docs underdocs/api/anddocs/developer/subsystems/for references to the file name before renaming.Underworld development team with AI support from Claude Code