[llvm] [llvm-readobj] Support --string-table for COFF (PR #141552)

[James Henderson via llvm-commits]

================
@@ -0,0 +1,60 @@
+# RUN: yaml2obj %s -o %t.obj
----------------
jh7370 wrote:

You're missing my point completely. I'm asking you to verify that the outward behaviour of llvm-readobj is correct through tests, not just reading the code. Putting it another way, it would not be impossible for somebody to decide that the COFFObjectFile interface is no longer the appropriate way to implement llvm-readobj for COFF and to implement something themselves that reads and prints COFF object information. In such a case, the tests should work for both the old and new implementation.

https://github.com/llvm/llvm-project/pull/141552