aokblast wrote:
> > > with the test limited to FreeBSD.
> >
> >
> > Is this because it assumes things about the call stack? That `vfprintf` is
> > its own function, I think?
> > If that's so could the test be changed to be two functions we write
> > ourselves, and step out of those?
>
> Yes, since step-out is to get the previous caller and we cannot assume
> different libc implementations have same call stack.
>
> I also try to create a testcase like the following:
>
> ```
> void g() {}
> void f() {g()}
> int main() {f()}
> ```
>
> and set a breakpoint on g(), step-out twice. It seems that simple program
> does not break lldb on FreeBSD.
One issue is that a userspace program can modify its memory mappings
dynamically (e.g., adding or removing mappings). On FreeBSD's case, the runtime
linker (rtld) may load libc after debugging has already started. At the moment,
I don’t have a good way to create a generic test case for this scenario.
https://github.com/llvm/llvm-project/pull/195809
_______________________________________________
lldb-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
