[PATCH] D61996: [llvm-objdump]Improve testing of some switches #2

[Fangrui Song via Phabricator via llvm-commits]

MaskRay added inline comments.


================
Comment at: test/tools/llvm-objdump/X86/Inputs/source-interleave.ll:17
+define i32 @main() #0 !dbg !14 {
+entry:
+  %retval = alloca i32, align 4
----------------
rupprecht wrote:
> MaskRay wrote:
> > I don't know how you feel about it, but I think these functions can just be empty. That is sufficient to emit debug info entries which will be used by the line table parser.
> I think having the real implementation (even if it's relatively short) will be a more realistic test of interleaved source. But maybe I don't understand your comment?
I ask because I find these irrelevant lines make the test hard to read. I am fine if you decide to keep the status quo.

What is important here is the source, the IR is just to provide debug information. If the function body is empty, we may argue that after optimizations the function is known to be empty.....




Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D61996/new/

https://reviews.llvm.org/D61996