Fix #15113 FP nullPointer with pthread functions

[chrchr-github]

No description provided.

[dmcppcheck]

AI review

I found no problems. Removing <not-null/> from argument 2 of pthread_cond_init and argument 2 of pthread_cleanup_push matches POSIX: a NULL attr means the default attributes, and arg is passed unchanged to the cleanup routine. <not-uninit/> is kept, so an uninitialised argument is still reported for both. Other functions that accept NULL, such as pthread_mutex_init and pthread_create, already allow it in the config, so the change matches them. The new tests in test/cfg/posix.c cover the NULL case and the uninitialised case.

Testing

• I built cppcheck from pr/. I did not build testrunner because this PR changes only cfg/ and test/cfg/.
• I ran the library test on test/cfg/posix.c with the options from test/cfg/runtests.sh (--check-library --check-level=exhaustive --enable=style,information --inconclusive --inline-suppr --debug-warnings --library=posix). It reported nothing, so there were no unmatched suppressions and no unexpected warnings.
• I compared the old and new config on a scratch file. I ran the PR's binary on it twice, once with --library=/tmp/baseposix.cfg (a copy of the base posix.cfg) and once with --library=posix:

#include <pthread.h>
void f(pthread_cond_t *c, void (*r)(void*)) {
    pthread_cond_init(c, NULL);
    pthread_cleanup_push(r, NULL);
    pthread_cleanup_pop(0);
    pthread_cond_t c2;
    pthread_cond_init(&c2, 0);
    pthread_condattr_t *a;
    pthread_cond_init(&c2, a);
    void *p;
    pthread_cleanup_push(r, p);
    pthread_cleanup_pop(1);
}

• Base config: 3 nullPointer false positives (lines 3, 4 and 7) and 2 uninitvar (lines 9 and 11).
• PR config: only the 2 correct uninitvar warnings (lines 9 and 11).

---

Automated review of a76d465 by Claude. The pull request was built and tested locally.