[PATCH] Fix for Bug 5053

[Reid Kleckner]

Looks pretty good. Do you want me to tweak the test case and commit this for you?

I checked the test case and it fails for me today, so this fixes a real bug.

================
Comment at: test/ExecutionEngine/win-frem.ll:1
@@ +1,2 @@
+; LLI.exe used to crash on Windows\X86 when certain single precession 
+; floating point intrinsics (defined as macros) are used.
----------------
These tests run on all platforms supported by the JIT, so I'd just name it "frem.ll". Windows just happened to be the problematic one.

================
Comment at: test/ExecutionEngine/win-frem.ll:5
@@ +4,3 @@
+;
+; RUN: lli %s > /dev/null
+
----------------
Use %lli, which expands with the appropriate -mtriple flag. You can also FileCheck the result.

================
Comment at: test/ExecutionEngine/win-frem.ll:7-8
@@ +6,4 @@
+
+target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32"
+target triple = "i686-pc-win32"
+
----------------
Get rid of this so we can be target neutral.

================
Comment at: test/ExecutionEngine/win-frem.ll:15
@@ +14,3 @@
+
+define i16 @main() {
+  %flt = load float* @flt
----------------
main returns i32?

http://reviews.llvm.org/D5387