From the 2026-08-02 full-repo review (finding F-07). Reproduced.
Repro: ./build/examples/recipe_registry_gen <out-dir> numsim_material (documented at README.md:98) → terminate called after throwing an instance of 'std::runtime_error' ("NumSimMaterialTarget: first increment supports exactly one scalar state variable...") when the catalogue reaches an unsupported recipe. Earlier files are left on disk; remaining recipes are never emitted. standalone and moose complete.
Fix: wrap the per-recipe emit in try/catch → print SKIPPED (<reason>) and continue; exit non-zero only if zero recipes succeeded. ~10 lines in examples/recipe_registry_gen.cpp.
Root cause note: there is no way to ask a target "can you emit this recipe?" other than catching std::runtime_error — see the companion Target::can_emit() issue. Related: #49.
From the 2026-08-02 full-repo review (finding F-07). Reproduced.
Repro:
./build/examples/recipe_registry_gen <out-dir> numsim_material(documented at README.md:98) →terminate called after throwing an instance of 'std::runtime_error'("NumSimMaterialTarget: first increment supports exactly one scalar state variable...") when the catalogue reaches an unsupported recipe. Earlier files are left on disk; remaining recipes are never emitted.standaloneandmoosecomplete.Fix: wrap the per-recipe emit in try/catch → print
SKIPPED (<reason>)and continue; exit non-zero only if zero recipes succeeded. ~10 lines inexamples/recipe_registry_gen.cpp.Root cause note: there is no way to ask a target "can you emit this recipe?" other than catching
std::runtime_error— see the companionTarget::can_emit()issue. Related: #49.