test command

Discover, build, and run Fortran test programs with a single command.

fobis test [options] [-- EXTRA_ARGS...]

How to use it, with examples: Testing.

All options

Option	Value	Default	Description
--fobos, -f	TEXT		Specify a "fobos" file named differently from "fobos"
--fci			Assume fobos inputs as case insensitive
--mode	TEXT		Select a mode defined into a fobos file
--shadow	TEXT		Rebuild the mode as a shadow (checks, sanitize or strict): the mode's settings with checking flags, in its own build directory (.fobis/shadow/<mode>-<shadow>/).
--compiler	TEXT	gnu	Compiler used (case insensitive, default gnu)
--suite	TEXT		Run only tests tagged with this suite name
--filter	TEXT		Run only tests whose name matches this glob pattern
--timeout	FLOAT		Per-test timeout in seconds [default: fobos [test] timeout, else 60]
--no-build			Skip (re-)compilation; run existing test binaries
--list			List discovered tests without running them
--coverage			Generate coverage report after running tests
--explain			When the program fails, explain why: from facts (location, declarations, debug mode to use, known compiler bugs), then with the LLM (cited cause and fix) when the location can be trusted.
--explain-context			Like --explain, but print what the model would receive instead.
--runner	TEXT	local	Where the programs run: local, or slurm (jobs submitted with sbatch; [scheduler] settings).
--launcher	TEXT		Start the program through this command: an MPI launcher, a debugger, a checker ('mpirun -np {ranks}', 'gdb --args', 'valgrind'); {ranks} is replaced by --ranks [default: mode key launcher]
--ranks	INTEGER		Number of MPI ranks: replaces {ranks} in the launcher (without a launcher: mpirun -np {ranks})
--jobs, -j	INTEGER		Test programs running at the same time (built first, one after the other); a test with --ranks takes as many [default: fobos [test] jobs, else 1]
--quiet, -q			Less verbose output
--json			Emit structured JSON to stdout instead of human-readable output (useful for scripting / agents)