src/error.c now scores 81.2%: 238 of 293 mutants killed, 204 by a
failing test, 24 by failing to compile, and 10 by hanging the suite --
deleting akerr_mutex_init() or the akerr_initializing re-entry guard
deadlocks the first test, which is the right answer for a broken lock.
Lock deletions are the one survivor category where surviving does not
mean harmless, so measure it rather than assume: rebuilt, the surviving
"delete the pool lock" mutant fails tests/err_threads_pool.c in 4 of 10
runs and fails under scripts/thread_test.sh in 5 of 5, with no false
positive on the unmutated library. The property assertions alone are a
coin flip on a missing lock; the sanitizer run is what holds that line.
The harness builds mutants with default CMake options and so never sees
it -- TODO item 8.
Also warn that a sanitized test binary run by hand does not inherit the
halt_on_error CTest gives it, and will print a race and still exit 0.
That is how the 5-of-5 above first read as 2 of 5.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every entry point may now be called from any thread. akerr_init() runs
exactly once however many threads race into it, the pool hands each slot
to exactly one thread, and reservations, registrations and lookups are
serialized against each other.
One recursive lock covers both tables (src/lock.h, private). Recursive
because raising an error re-enters the library -- FAIL needs a pool slot
and a status name -- and single because two locks would mean an ordering
to get wrong. Registry bodies that use the early-returning FAIL_*_RETURN
macros are split into *_locked functions behind wrappers that take and
release the lock on one path; consumer callbacks are never called under
it.
This is an ABI break, hence 2.0.0 and SOVERSION 2:
- akerr_next_error() now returns a context that already holds its
reference. Finding a free slot and claiming it has to be one operation,
or two threads scanning at once are handed the same slot.
ENSURE_ERROR_READY no longer increments.
- __akerr_last_ignored is thread-local, as is the last-ditch context used
to report akerr_release_error(NULL).
The threading backend is chosen at configure time by AKERR_THREADS
(auto, pthread, none). auto fails the configure when it cannot find
POSIX threads rather than quietly building a library that reports itself
thread safe and is not. generrno.sh stamps the decision into the
generated header as AKERR_THREAD_SAFE, so a consumer cannot disagree
with the library about it.
Tests: err_threads_init, err_threads_pool and err_threads_registry
assert exclusive slot ownership, exactly one winner for a contested
range, and every registered name readable back under contention.
AKERR_SANITIZE builds the library and the tests with any sanitizer;
scripts/thread_test.sh runs the suite under ThreadSanitizer and CI runs
it. Removing the pool lock makes both the sanitizer and the plain
assertions fail, so the tests are not vacuous.
Documented in README.md and UPGRADING.md, including what this does not
cover: renaming a status while another thread looks it up, and which of
two simultaneous unhandled errors sets the exit status.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
scripts/coverage.py configures an instrumented build tree
(-DAKERR_COVERAGE=ON), runs the CTest suite in it, and reports merged
gcov line/branch/function coverage per library source. Like the mutation
harness it has no third-party dependencies and supports --threshold and
--junit; thresholds gate each file as well as the total so the generated
status-name table cannot mask a regression in src/error.c.
Only the library is instrumented. The public header's macros cannot be
measured this way -- GCC attributes an expanded macro to its call site,
so header logic would report as lines of the test that used it -- which
is what mutation testing against include/akerror.tmpl.h is for.
Coverage flags are applied per target rather than globally, so they do
not leak into the exported/installed target interface.
Current numbers for src/error.c are 94.0% line and 59.5% branch; the CI
gate is set to 90/50 to keep headroom, matching the convention used for
the mutation score threshold.
Tests run: ctest (23/23), cmake --build build --target coverage,
threshold gate verified failing at --threshold 99, cmake --install
checked for flag leakage, out-of-tree --build-dir checked.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
- Added explicit status validation across error handling tests.
- Added lifecycle and slot-leak checks to older tests.
- Improved unhandled-error propagation coverage.
- Added repository guidance and a test runner script.
- Verified all 23 CTest tests pass.
Co-Authored by Codex GPT 5.4
