[PATCH] D60311: MIR printer should lowercase sub-register names to be in sync with parser?

[Markus Lavin via Phabricator via llvm-commits]

markus added a comment.

In D60311#1456007 <https://reviews.llvm.org/D60311#1456007>, @bjope wrote:

> I think it would be better to let the printer use the same convention as the specified in the definitions (as used in the code).
>  And then I think it is a bug in the parser, as it tries to be case insensitive by matching the parsed string against a lower case definition. It should also lower the parsed string if we want the parser to be case insensitive.
>
> If we fix the parser instead, then we still would get debug printouts etc that matches the mixed case used in the C++ code. Or maybe I haven't understood the problem correctly.


I agree that this should be fixed in the parser. All that is needed is to decide if it should be case sensitive or not. 
To me case sensitivity makes most sense as the instruction and register names used could be rather complex for certain targets.


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D60311/new/

https://reviews.llvm.org/D60311