[Lldb-commits] [lldb] [lldb] Use correct size when dumping DWARF64 DW_FORM_ref_addr (PR #146686)

[David Spickett via lldb-commits]

================
@@ -416,12 +416,8 @@ void DWARFFormValue::Dump(Stream &s) const {
   case DW_FORM_ref_addr: {
     assert(m_unit); // Unit must be valid for DW_FORM_ref_addr objects or we
                     // will get this wrong
-    if (m_unit->GetVersion() <= 2)
-      DumpAddress(s.AsRawOstream(), uvalue, sizeof(uint64_t) * 2);
-    else
-      DumpAddress(s.AsRawOstream(), uvalue,
-                  4 * 2); // 4 for DWARF32, 8 for DWARF64, but we don't
----------------
DavidSpickett wrote:

I wouldn't be surprised if it was just incorrect forever and no one has used this code in a long time. Generally dump() is used by llvm developers when debugging.

I doubt we want to add tests for this, but I put this fix in review just in case people think we should.

https://github.com/llvm/llvm-project/pull/146686