LLDB doesn’t stop on SIGABRT (assert) with x86 executable on Apple ARM

How can this be fixed?

I recommend that you do what Jim suggested on that thread you referenced [1], namely file a bug.

Please post your bug number, just for the record.

In terms of workarounds, if you modify your assert macro to not use abort but rather invoke __builtin_trap(), that seems to make LLDB much happier.

If you can’t modify your assert macro then you could try something like this:

#include <stdio.h>
#include <stdlib.h>

static void myAbort(int sigNum) {
    __builtin_trap();
}

int main(int argc, const char * argv[]) {
    signal(SIGABRT, myAbort);
    fprintf(stderr, "Hello Cruel World!\n");
    abort();
    fprintf(stderr, "Goodbye Cruel World!\n");
    return 0;
}

This does really weird stuff when you run it directly [2], but it seems to yield reasonable results when debugging with LLDB.

btw I’m testing on macOS 14.0 with Xcode 15.0. YMMV on other configs.

Share and Enjoy
—
Quinn “The Eskimo!” @ Developer Technical Support @ Apple
let myEmail = "eskimo" + "1" + "@" + "apple.com"

[1] https://discourse.llvm.org/t/lldb-doesnt-stop-on-sigabrt-assert-with-x86-executable-on-apple-arm/69749/1

[2] I won’t hold that against Rosetta. Honestly, I’m in awe of its ability to make any of this stuff work.

Thanks for the reply.

I recommend that you do what Jim suggested on that thread you referenced [1], namely file a bug. Please post your bug number, just for the record.

I have immediately reported the bug. Here is the number. FB13382646

In terms of workarounds, if you modify your assert macro to not use abort but rather invoke __builtin_trap(), that seems to make LLDB much happier.

I tried __builtin_trap. As a result, LLDB stopped when the app crashed. Thank you very much.