clayborg added a comment. I see you moved the JIT code over into ClangUserExpression. I was wondering if all languages might not want to generate llvm IR and then let the JIT make code from the llvm IR? This is the reason were originally left the IR in the UserExpression. IR is very generic and can be applied to just about anything. Any interest in using llvm IR for the Go expression parser? I am assuming you have made a expression tree that you evaluate. This would be very similar to what we do. When we evaluate an expression currently we generate the llvm IR and then we say "that is easy to emulate (just a bunch of loads and stores), so we don't actually need to JIT it up, we can just emulate the IR. If it is too complex, we pass this along to the JIT and let it generate code for us.
Repository: rL LLVM http://reviews.llvm.org/D13073 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
