verificationCases/README.md
Verification cases
These programs compare the implemented model with analytical solutions or exact geometric integrals. Experimental film-thickness comparisons belong in the Newtonian validation report.
| Case | Check |
|---|---|
embedAxiVofAdvection.c |
Prescribed axisymmetric interface advection |
laplaceEmbedTube.c |
Young–Laplace pressure jump on a sequence of uniform grids |
laplaceEmbedTubeAdapt.c |
Young–Laplace pressure jump and coupling checks on an adaptive grid |
staticFilmTube.c |
Static interface near the wall, with density and viscosity contrasts |
tubeGeometry.c |
Exact tube volume, axial area and constant-flow flux residual through repeated adaptation |
Use the project-local Basilisk installation
configured by .project_config. From the
repository root:
bash runTests.sh
bash runTests.sh --verification
bash verificationCases/runVerification.sh verificationCases/tubeGeometry.c
VERIFICATION_THREADS=16 bash runTests.shBuilds and diagnostic output are written to ignored
build-<case>/ directories. Each
verification program returns nonzero on failure and
prints a bare PASS or FAIL
line. Record the toolchain and thread count when
retaining results.
The driver also builds
laplaceEmbedTubeAdapt.c with
-DSKIP_EMBED_GUARDS. This negative control
uses active-cell metric initialisation without the
complete stored-tree tube reconstruction, then omits the
post-adaptation sync and cleanup calls. It must fail a
consistency probe. It tests the combined compatibility
pathway rather than isolating the contribution of each
call.