Files
libakerror/tests
Andrew Kesterson 076b80c846
Some checks failed
libakerror CI Build / cmake_build (push) Successful in 2m48s
libakerror CI Build / coverage (push) Successful in 2m47s
libakerror CI Build / thread_sanitizer (push) Failing after 2m50s
libakerror CI Build / mutation_test (push) Successful in 39m13s
Record the mutation score after the exit-status fix
src/error.c now scores 81.4%: 245 of 301 mutants killed, 211 by a
failing test, 24 by failing to compile, and 10 by hanging the suite.
The population grew by 8 with akerr_exit(), and all 8 die, as do the 4
in akerr_default_handler_unhandled_error() -- no survivor anywhere in
either function.

Two of those are worth naming. Mutating the guard to status < 1 (the
sentinel-for-zero behaviour this library deliberately does not have) and
mutating the NULL-context exit(1) to exit(0) both die, so the tests pin
the two ways an exit could start claiming success rather than just the
one that was broken.

That moves the default handler out of the "needs a subprocess test"
survivor category noted here: it has one now. The default *logger* is
still in it -- every other test swaps in the capturing logger, so
nothing watches what the real one writes to stderr.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-01 07:57:33 -04:00
..
2026-07-29 17:12:52 -04:00
2026-07-29 17:12:52 -04:00
2026-07-29 17:12:52 -04:00