[Lldb-commits] [PATCH] D98996: Teach DWARFExpression about DWARF 4+ Location Descriptions

Mon Mar 22 01:06:04 PDT 2021

labath added a comment.

Are you sure that this logic is correct in presence of DW_OP(_bit)_piece? If I follow this right, then the final value type will be determined by the last operand. That sounds like it could be right for regular dwarf expressions, but I'm not sure about those with pieces. The classification function will mark those as "memory", and that doesn't sound right...

================
Comment at: lldb/unittests/Expression/DWARFExpressionTest.cpp:360-361
                        llvm::HasValue(Scalar(LLDB_INVALID_ADDRESS)));
+  // Memory location: *(*0x4).
+  // Evaluate returns LLDB_INVALID_ADDRESS for all load addresses.
+  EXPECT_THAT_EXPECTED(Evaluate({DW_OP_lit4}, {}, {}, &exe_ctx),
----------------
This comment doesn't seem right.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98996/new/

https://reviews.llvm.org/D98996