Hi again
OK, a few more details.
There is nothing that you can do about SIGKILL. Your process gets terminated
and you get no chance to do anything about it. Valgrind cannot do it’s final
resource leak checks and your log is probably truncated.
SIGTERM, on the other hand, gives you a chance to do something - handle it or
ignore it. You could write some code like this
#include “valgrind.h”
static void handle_terminate_valgrind(int signo)
{
/* exit cleanly */
/*
* if not possible to exit cleanly
* use VALGRIND_DO_LEAK_CHECK from memcheck.h
*/
}
/* somewhere In your startup code */
if (RUNNING_ON_VALGRIND) {
signal(SIGTERM, handle_terminate_valgrind);
/* alternatively to ignore
signalSIGTERM, SIG_IGN);
/*
}
A+
Paul
_______________________________________________
Valgrind-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/valgrind-users
