Ticket #19 (defect)
Opened 1 year ago
Last modified 1 year ago
pointer_to_shared_neg should not be a negative test
Status: closed (duplicate)
| Reported by: | bonachea@cs.berkeley.edu | Assigned to: | somebody |
|---|---|---|---|
| Priority: | major | Milestone: | |
| Component: | component1 | Version: | |
| Keywords: | Cc: | ||
pointer_to_shared_neg.upc:
shared int *s1;
shared int control_int_variable;
int main()
{
/*
* To check that pointers to shared objects cannot be cast to
* pointers to local objects, provided that the shared object
* does not have affinity to the executing thread
* the compiler MUST raise an ERROR here !!!
*/
int *p1;
if(MYTHREAD == 0) {
control_int_variable = 55;
s1 = &control_int_variable;
p1 = (int*) s1;
}
upc_barrier;
return 0;
}
Despite the comment, this code is actually fully correct and should not generate a failure at compile or run time.
If the intent was to MYTHREAD == 1, the code would be erroneous but no compile or runtime diagnostic is required by the spec (see ticket #18).
Change History
09/25/07 18:14:59: Modified by apokayi
- status changed from new to closed.
- resolution set to duplicate.
