Skip to content

Testing and CI

How the driver is checked: host tests against a model that shares no code with it, a fuzzer that controls what the ADC returns, compiled documentation, an install test, Cortex-M and ESP32 builds, static analysis and measured quality gates, on every pull request.

Host tests

cmake -S . -B build -DNTC_THERMISTOR_BUILD_TESTS=ON
cmake --build build
ctest --test-dir build --output-on-failure
File Covers
tests/test_ntc_thermistor.cpp Validate() at every boundary (as static_assert), presets as data, lifecycle, the rails and margins in both topologies, the window and offset, averaging, Unstable, the bound, the helpers
tests/test_model.cpp The driver against the model: Ln16(), exactness at the rows, agreement with the true curve of every part over its range, interpolation bounds, round trip, symmetry, separate supply and reference, no wrap, purity
tests/test_curves.cpp Steinhart-Hart mode against std::log in both topologies and its validation, the bound's terms (B on a table, curve_error_mc), the faults with a reference above the supply, the table ends, the cap on the waits
tests/generated/ The output of tools/ntc_table.py (built at test time) is a valid table that Convert() follows to 0.06 K; the tool's own checks run as a test
tests/test_bias.cpp The order of port calls and the divider off on every path
tests/test_errors.cpp Every error value the driver can return is returned by a real call; ToString, Result, version
tests/test_kconfig.cpp The menuconfig defaults, the curve choices
tests/mock_ntc_thermistor.hpp The model and the mock ports
tests/compile_fail/ Ports and configurations that must not compile
tests/check_install.cmake find_package(ntc_thermistor) from an installed package
tests/docs/ Every code example on the docs pages, compiled

The tests build with -fno-exceptions -fno-rtti and warnings as errors, on GCC and Clang.

The model and the parts

The model is written independently of the driver: floating point (long double), a divider and ADC model, and the resistance law of each part, with no code from inc/. It covers seven parts; the TDK preset is checked against the model's separately typed copy of its rows (P7), and tables are also derived from the beta parts at 10, 5 and 2.5 degree steps for the step rule:

Part Law Driver gets Why
P1 Beta 3950, 10 kOhm R25 and B The generic class
P2 Beta 4250, 100 kOhm R25 and B A high R25
P3 Beta 3100, 2.2 kOhm R25 and B A low R25, hot resistances of a few ohms
P4 Steinhart-Hart, 47 kOhm A table, 2.5 degree rows A curve that is not beta
P5 Steinhart-Hart, 10 kOhm A table, uneven rows Coarse and uneven spacing
P6 Beta 3380, 10 kOhm at 20 degrees R25 and B A reference other than 25 degrees
P7 The example part's rows, typed again A table A table that is data, not a formula

Each case runs for the parts and for both divider topologies, 10, 12 and 16 bits. The interpolation bounds the tests assert are pinned numbers measured from the model (see Conversion), not loosened to pass.

Fuzz tests

Everything the driver reads is a code from the port's ADC, so the code is the untrusted input. tests/fuzz/fuzz_ntc_thermistor.cpp lets libFuzzer choose a configuration (a beta curve, the example preset, the generic preset, a table built from fuzzed pairs, or none), then up to 48 driver calls, and on every conversion what the ADC does: pass the code the divider model gives for the simulated resistance, fail, return a rail code, a value above scale, or a code that wobbles. After every call it checks the driver's promises and aborts, naming the line, when one breaks:

  • a failed call carries an Error this driver returns and a zeroed value; a node code at the open rail is never Ok and never the other fault;
  • before EnsureInitialized() succeeds no port call is made; InvalidArgument means none was made; an empty curve, or a beta curve without a window, is always InvalidArgument;
  • an Ok temperature lies inside the window and, in table mode, inside the table's span plus the offset; in beta mode the temperature is above absolute zero before the offset;
  • Ok never comes from a failed sample, a sample at or above 2^bits, or a spread above max_spread_codes; uncertainty_mc is at least 1 and the mean code is the rounded mean;
  • SetBias is switched on and off the same number of times in a read and the last call is off, on every error path;
  • over 64 sorted probe codes Convert is monotone and its errors come as contiguous runs in order;
  • every ADC call has channel == Config::adc_channel, a read makes at most sample_count conversions, and a bad SetConfig() leaves the old configuration.
# libFuzzer needs Clang (on macOS, Homebrew's LLVM)
cmake -S . -B build-fuzz -DCMAKE_CXX_COMPILER=clang++ -DNTC_THERMISTOR_BUILD_TESTS=ON -DNTC_THERMISTOR_BUILD_FUZZERS=ON
cmake --build build-fuzz --target fuzz_ntc_thermistor
./build-fuzz/tests/fuzz/fuzz_ntc_thermistor -max_total_time=120 tests/fuzz/corpus/fuzz_ntc_thermistor

The same file also builds without libFuzzer as fuzz_replay_fuzz_ntc_thermistor, and CTest replays every input in tests/fuzz/corpus/fuzz_ntc_thermistor/ on GCC and Clang (ntc_thermistor_fuzz_ntc_thermistor_corpus): the seeds, and every input that ever broke a promise (name it regress-*), so a fixed bug stays fixed. CI fuzzes for 120 s on every pull request, from the seeds plus a corpus it keeps between runs (job Host quality / fuzz).

Quality gates

The Host quality job measures the library code, inc/ and src/ only (tests and examples are left out), and fails when a value is over the limits of this repository's assurance level, high (set in .kokeb/handbook.yml, defined by the handbook's ADR-0020):

Gate Tool and how it is measured Limit at high
Complexity lizard 1.24.1, modified cyclomatic complexity (-m: a switch counts once) per function CCN at most 15
Function length lizard, lines of code per function at most 60
Parameters lizard, per function at most 6
Line coverage GCC 14 --coverage -O0, every CTest, gcovr 8.7 --merge-lines (a template's lines count once across its instances), --exclude-unreachable-branches --exclude-throw-branches at least 85 %
Branch coverage same run at least 75 %
MC/DC GCC -fcondition-coverage, same run measured, not gated at high (100 % at safety)
Sanitizers Clang, -fsanitize=address,undefined -fno-sanitize-recover=all, every CTest no report
Fuzzing libFuzzer with ASan and UBSan, 120 s per target no broken promise

The reports are in the run's host-quality-reports artifact and its step summary. To run the gates locally:

# Coverage (GCC 14 or later), then the complexity limits
cmake -S . -B build-cov -DCMAKE_CXX_COMPILER=g++-14 -DNTC_THERMISTOR_BUILD_TESTS=ON \
  -DCMAKE_CXX_FLAGS="--coverage -O0 -fcondition-coverage"
cmake --build build-cov && ctest --test-dir build-cov
gcovr --root . --filter inc/ --filter src/ --merge-lines --exclude-unreachable-branches --txt build-cov
lizard -m -C 15 -L 60 -a 6 inc src

# Sanitizers (Clang)
cmake -S . -B build-san -DCMAKE_CXX_COMPILER=clang++ -DNTC_THERMISTOR_BUILD_TESTS=ON \
  -DCMAKE_CXX_FLAGS="-fsanitize=address,undefined -fno-sanitize-recover=all -O1 -g"
cmake --build build-san && ctest --test-dir build-san --output-on-failure

The fuzz gate runs locally as shown in Fuzz tests.

clang-tidy 20 (_config/.clang-tidy: bugprone, CERT, clang-analyzer, cppcoreguidelines, modernize, performance, portability and readability, each disabled check with its reason) and cppcheck (strict) run in the C++ lint and Repository checks jobs and fail on any finding.

The MISRA job (ci-misra.yml) runs the shared ru-misra workflow of ci-tools at level high: it traces a host build and the Cortex-M build (ntc_thermistor_optimised_build) of every translation unit, checks them against MISRA C++:2023 and the handbook's C++20 profile (the CodeQL coding-standards and profile packs, plus the validated clang-tidy and cppcheck checks), and reports each finding against the handbook's compliance matrix and deviation records. A finding is fixed in the code; a deliberate exception names its deviation on the line before it (// codeql::misra_deviation_next_line(MD-CPP-009) above the .ipp include of inc/ntc_thermistor.hpp). Until the repository is public the job runs with open_source_prerelease: true, the declaration that the code is released under the licence in LICENSE and is to be published.

CI workflows

Everything runs from ci.yml, which calls one workflow file per check and ends in the CI gate job, the single required check (CI gate / gate).

Job What it checks
Host tests GCC, then Clang (one job), build and run every test above, then build and run examples/host on its own
Host quality The gates above: complexity, coverage, sanitizers, fuzzing
Cross-compile arm-none-eabi-g++ for Cortex-M0+, M4F and M7F with the strict flags, warnings as errors (nothing linked)
C++ lint clang-format and clang-tidy over the sources
MISRA MISRA C++:2023 and the C++20 profile on traced host and Cortex-M builds, at level high (see Quality gates)
Docs Strict site build (_config/mkdocs.yml), generated-region drift, Doxygen (_config/Doxyfile), offline link check
Repository checks One job: DCO on pull requests, the org standard, Markdown and YAML lint, the Kokeb Labs handbook rules for .kokeb/handbook.yml (by rule ID), template drift (warns only) and cppcheck over the driver sources
CI gate Passes only if every job above passed: the one required check

The ESP32 examples job builds every app in examples/esp32/app_config.yml with the repository root as the driver's component. The gate needs it too.