[Lldb-commits] [lldb] r188124 - Fixed a case where GCC was emitting a DW_TAG_class_type that has a DW_AT_declaration set to true, yet the class actually contains a definition for the class in that DIE.

[Malea, Daniel]

Hi Greg,

This commit seems to introduce a few errors on the buildbots in the following tests:
- TestDynamicValue
- TestVirtual

http://lab.llvm.org:8011/builders/lldb-x86_64-linux/builds/6709

When you have a moment, can you take a look?


Thanks,
Dan

-----Original Message-----
From: lldb-commits-bounces at cs.uiuc.edu [mailto:lldb-commits-bounces at cs.uiuc.edu] On Behalf Of Greg Clayton
Sent: Friday, August 9, 2013 8:10 PM
To: lldb-commits at cs.uiuc.edu
Subject: [Lldb-commits] [lldb] r188124 - Fixed a case where GCC was emitting a DW_TAG_class_type that has a DW_AT_declaration set to true, yet the class actually contains a definition for the class in that DIE.

Author: gclayton
Date: Fri Aug  9 19:09:35 2013
New Revision: 188124

URL: http://llvm.org/viewvc/llvm-project?rev=188124&view=rev
Log:
Fixed a case where GCC was emitting a DW_TAG_class_type that has a DW_AT_declaration set to true, yet the class actually contains a definition for the class in that DIE.


Modified:
    lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp

Modified: lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp
URL: http://llvm.org/viewvc/llvm-project/lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp?rev=188124&r1=188123&r2=188124&view=diff
==============================================================================
--- lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp (original)
+++ lldb/trunk/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp Fri Aug  9 19:09:35 2013
@@ -6193,6 +6193,30 @@ SymbolFileDWARF::ParseType (const Symbol
                     GetUniqueDWARFASTTypeMap().Insert (type_name_const_str, 
                                                        unique_ast_entry);
                     
+                    if (is_forward_declaration && die->HasChildren())
+                    {
+                        // Check to see if the DIE actually has a definition, some version of GCC will
+                        // emit DIEs with DW_AT_declaration set to true, but yet still have subprogram,
+                        // members, or inheritance, so we can't trust it
+                        const DWARFDebugInfoEntry *child_die = die->GetFirstChild();
+                        while (child_die)
+                        {
+                            switch (child_die->Tag())
+                            {
+                                case DW_TAG_inheritance:
+                                case DW_TAG_subprogram:
+                                case DW_TAG_member:
+                                case DW_TAG_APPLE_property:
+                                    child_die = NULL;
+                                    is_forward_declaration = false;
+                                    break;
+                                default:
+                                    child_die = child_die->GetSibling();
+                                    break;
+                            }
+                        }
+                    }
+
                     if (!is_forward_declaration)
                     {
                         // Always start the definition for a class type so that


_______________________________________________
lldb-commits mailing list
lldb-commits at cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits