[Lldb-commits] [PATCH] D11119: Parse breakpoint expressions using the language of the frame's CU.

Dawn Perchik dawn+llvm at burble.org
Wed Jul 15 11:43:47 PDT 2015

dawn added a comment.

In http://reviews.llvm.org/D11119#205491, @jingham wrote:

> That sounds fine, except you shouldn't call the option "expr-language" because [...]
>  Since you are only going to add the language options to versions of breakpoint set that take some form of identifier (--name, --method --base-name) maybe --identifier-language would be better?

Actually, I do plan for this option to apply to the expression in '--address <expr>' as well.  The

  settings set target.expr-language <lang>

is intended to affect how expressions used in all commands are parsed and interpreted.  This patch is just part of that bigger plan, by addressing how the name_type_masks are set.

We want the command options to match the target's setting for consistency.   For example:

  expr --expr-language <lang> <expr>
  breakpoint set --expr-language <lang> <expr>

So '--identifier-language' isn't quite right, and my concern with plain '--language' is that

  settings set target.language <lang>

might have users thinking it affects everything from unmangling to how the DWARF is read.

'--expr-language' seemed to to fit best to me.  Would you agree?



More information about the lldb-commits mailing list