1
0
mirror of https://github.com/ioacademy-jikim/debugging synced 2025-06-10 17:36:21 +00:00
debugging/03_day/check-0.10.0/tests/check_nofork_teardown.c
2015-12-13 22:34:58 +09:00

63 lines
1.6 KiB
C

#include "../lib/libcompat.h"
#include <stdio.h>
#include <stdlib.h>
#include "check.h"
/**
* This test checks the result if in CK_NOFORK
* mode a unit test fails but a checked teardown
* runs after the failed test.
*
* Previously, the failure would be reported as:
*
* 0%: Checks: 1, Failures: 1, Errors: 0
* (null):-1:S:tc:will_fail:0: Assertion '0' failed
*
* The reason why this happens is this: the end of the
* message sequence coming down the pipe is CK_MSG_LOC
* (location of failing test), CK_MSG_FAIL, CK_MSG_CTX
* (TEARDOWN). It is this final message that confuses
* things, because rcvmsg_update_ctx() updates
* rmsg->lastctx (which likely is the right thing for it
* to do), which is the ctx value used by the first 'if'
* body in construct_test_result() in its call to
* tr_set_loc_by_ctx().
*
* After the fix, the test failure should be reported
* as:
*
* 0%: Checks: 1, Failures: 1, Errors: 0
* check_nofork_teardown.c:33:F:tc:will_fail:0: Assertion '0' failed
*/
START_TEST( will_fail )
{
ck_assert(0);
}
END_TEST
static void empty_checked_teardown( void )
{
}
int main( void )
{
SRunner *sr = srunner_create( NULL );
Suite *s = suite_create( "bug-99" );
TCase *tc = tcase_create( "tc" );
int result;
srunner_add_suite( sr, s );
srunner_set_fork_status( sr, CK_NOFORK );
suite_add_tcase( s, tc );
tcase_add_checked_fixture( tc, NULL, empty_checked_teardown );
tcase_add_test( tc, will_fail );
srunner_run_all( sr, CK_ENV );
result = srunner_ntests_failed( sr ) ? EXIT_FAILURE : EXIT_SUCCESS;
srunner_free( sr );
return result;
}