Open 5c4lar opened 2 years ago
This seems like a bug to me.
Really, I'm not sure why we track Type
as part of the symbolic_operand
relation at all. Most consumers of symbolic_operand
ignore it, and those that do use it should be able to trivially check code(Dest)
or !code(Dest)
.
Some
symbolic_operand
will be marked as bothcode
anddata
. This can be observed in almost every test case. Is there any reason for this ambiguity?