On Aug 3, 2011, at 2:32 PM, Bill Wendling wrote:
> This pains me to ask, because I'm the original author of this pass, but
is the LowerSetJmp pass dead? I created a simple program, manually converted it
to use the @llvm.setjmp/@llvm.longjmp intrinsics, ran "opt
-lowersetjmp" on it, and tried to compile it only to get this:
I think that the CBE still uses it, but it can't use it in a successful way.
I'm in favor of zapping it. Please mention it in the release notes.
-Chris
>
> Undefined symbols for architecture x86_64:
> "___llvm_sjljeh_throw_longjmp", referenced from:
> foo() in ccf8Gf0w.o
> "___llvm_sjljeh_init_setjmpmap", referenced from:
> _main in ccf8Gf0w.o
> "___llvm_sjljeh_add_setjmp_to_map", referenced from:
> _main in ccf8Gf0w.o
> "___llvm_sjljeh_destroy_setjmpmap", referenced from:
> _main in ccf8Gf0w.o
> "___llvm_sjljeh_is_longjmp_exception", referenced from:
> _main in ccf8Gf0w.o
> "___llvm_sjljeh_get_longjmp_value", referenced from:
> _main in ccf8Gf0w.o
> "___llvm_sjljeh_try_catching_longjmp_exception", referenced
from:
> _main in ccf8Gf0w.o
>
> A quick scan through the sources showed none of these symbols in it.
>
> Should we just get rid of this pass and the intrinsics?
>
> -bw
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev