Hi folks,
I was optimizing my code which uses gwlib and have hit a big problem
when I was converting things like
mutex_lock(my_list_mutex);
do_someting_with(my_list);
mutex_unlock(my_list_mutex);
to:
gwlist_lock(my_list);
do_someting_with(my_list);
gwlist_unlock(my_list);
to get rid of an additional mutex we dont really need as we have a
built in one in List.
While doing that I was hitting a deadlock issue.
It turned out that in my do_something part I called gwlist_sort and
that looks like this:
void gwlist_sort(List *list, int(*cmp)(const void *, const void *))
{
gw_assert(list != NULL && cmp != NULL);
gwlist_lock(list);
if (list->len == 0) {
/* nothing to sort */
gwlist_unlock(list);
return;
}
qsort(&GET(list, 0), list->len, sizeof(void*), cmp);
gwlist_unlock(list);
}
In my eyes, this function should call lock()/unlock(). not gwlist_lock
()/gwlist_unlock();
It is the only place where gwlist_lock is used in list.c.
Is this simply a bug I have spotted or was there anywhere a specific
reason why a sort uses the permanent lock instead of the
single_operation_lock?
Andreas Fink
Fink Consulting GmbH
---------------------------------------------------------------
Tel: +41-61-6666332 Fax: +41-61-6666331 Mobile: +41-79-2457333
Address: Clarastrasse 3, 4058 Basel, Switzerland
E-Mail: [EMAIL PROTECTED]
Homepage: http://www.finkconsulting.com
---------------------------------------------------------------
ICQ: 8239353
MSN: [EMAIL PROTECTED] AIM: smsrelay Skype: andreasfink
Yahoo: finkconsulting SMS: +41792457333
