Post

Replies

Boosts

Views

Activity

Trace/BPT trap in very simple C code compiled with clang
I wonder if this is correct behavior. I was surprised to get this result when compiling and running the following C code with Apple clang version 14.0.0 (clang-1400.0.29.102) target arm64-apple-darwin21.6.0 on a M1 Pro 12.7.6 with cc -O2 file.c: #include <stdio.h> #include <stdlib.h> unsigned long long factorial(int n) { unsigned long long fac = 1; while (n > 0) fac *= n; return fac; } int main() { return factorial(1); } Compiling with -O2 and running this code gives "Trace/BPT trap". Checking with LLDB: $ lldb ./a.out (lldb) target create "./a.out" Current executable set to '/Users/engelen/Projects/Euler/a.out' (arm64). (lldb) run Process 79580 launched: '/Users/engelen/Projects/Euler/a.out' (arm64) Process 79580 stopped * thread #1, queue = 'com.apple.main-thread', stop reason = EXC_BREAKPOINT (code=1, subcode=0x100003fb4) frame #0: 0x0000000100003fb4 a.out`main at 20.c:9:3 [opt] 6 unsigned long long fac = 1; 7 while (n > 0) 8 fac *= n; -> 9 return fac; 10 } 11 12 int main() The loop is non-terminating. But a breakpoint trap is triggered at the return statement. The code should just hang in the loop IMO, not trap, because it never updates variable n (a correct factorial function should decrement n). Never seen this before (not since I started wiring C code in the 80s.) If I change the update *= into += then there is no trap.
1
0
216
2w