[Lldb-commits] [Bug 12232] New: assert(ivar_decl != NULL) triggers in SymbolFileDWARF::ParseChildMembers for C++

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Sat Mar 10 11:25:08 PST 2012


             Bug #: 12232
           Summary: assert(ivar_decl != NULL) triggers in
                    SymbolFileDWARF::ParseChildMembers for C++
           Product: lldb
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Keywords: crash-on-invalid
          Severity: normal
          Priority: P
         Component: All Bugs
        AssignedTo: lldb-commits at cs.uiuc.edu
        ReportedBy: abigagli at gmail.com
    Classification: Unclassified

On OSX 10.7.3, LLDB built from trunk at 152495.

I have a "pure" C++ program (i.e. neither Objective-C nor Objective-C++) built
with fsf-gcc-4.6.2 in std=c++0x mode.
While debugging it with LLDB, i frequently hit the 

assert (ivar_decl != NULL);

at line 1645 in SymbolFileDWARF::ParseChildMembers.

The code around is as follows:

if (prop_name != NULL)

    clang::ObjCIvarDecl *ivar_decl =
    assert (ivar_decl != NULL);

    GetClangASTContext().AddObjCClassProperty (class_clang_type,

I verified (debugging LLDB itself) that prop_name has indeed a value.
But of course the dyn_cast to clang::ObjCIvarDecl fails (there is no
ObjCIvarDecl in sight here) resulting in ivar_decl being NULL and hence the
assert kicks in.

I'm not good enough to understand why this code gets under the impression
there's some ObjC properties, and at the moment I've applied the following
patch to be able to keep debugging: 
Index: SymbolFileDWARF.cpp
--- SymbolFileDWARF.cpp    (revision 152495)
+++ SymbolFileDWARF.cpp    (working copy)
@@ -1642,9 +1642,9 @@

                             clang::ObjCIvarDecl *ivar_decl =
-                            assert (ivar_decl != NULL);
+                            //assert (ivar_decl != NULL);

+                            if (ivar_decl)

That seems to be working for now (I'm able to continue stepping through over
the instructions that were previously asserting), but I'd rather prefer this to
be checked and possibly solved in a more substantial way.

Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

More information about the lldb-commits mailing list