[PATCH] D56223: [WebAssembly] Made InstPrinter more robust

[Heejin Ahn via Phabricator via llvm-commits]

aheejin added inline comments.


================
Comment at: lib/Target/WebAssembly/InstPrinter/WebAssemblyInstPrinter.cpp:304
+  default:
+    return "invalid_type";
   }
----------------
aardappel wrote:
> aheejin wrote:
> > So does this print "invalid_type" even when we have an invalid instruction within itself?
> I'm not sure what you mean. It currently prints this only when an instruction explicitly tries to print an operand that is a block_type or other signature, and it happens to be an unknown value.
Oh what I meant was.. I was not sure if printing instructions like
```
block invalid_type
```
was ok. 
```
br 32 # invalid depth argument!
```
seemed ok because even if it's out of context, `br 32` itself is a valid instruction, and we are giving hints that this instruction does not make sense within the current context in the comment anyway. But `block invalid_type` is not an instruction and cannot be parsed. So I think in this case it deserves an assertion, but I don't feel very strongly about it. What do you think?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D56223/new/

https://reviews.llvm.org/D56223