[patch] PR 12028 "GDB crashes on a double free during overload resolution"

[sami wagiaalla]

On 09/22/2010 03:19 PM, Tom Tromey wrote:
>>>>>> "Sami" == sami wagiaalla<swagiaal@redhat.com>  writes:
>
> Sami>  I agree. I have attached a patch which has been regression tested on
> Sami>  x8664 with gcc-4.4.4-10.fc13
>
> This patch is also ok, assuming that it is still valgrind-clean for the
> test in question.
>

I ran valgrind like this:

valgrind ./gdb/gdb ./gdb/testsuite/gdb.cp/pr12028 -ex 'sta' -ex 'p 
D::foo(b)' -ex 'q'

There are no glaring errors. The original double free error is still 
gone of course and there are no others.

Incidentally, there is this error:

==8467== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 27 from 9)
==8467== Use of uninitialised value of size 8
==8467==    at 0x57C966: bcache_full (bcache.c:189)
==8467==    by 0x52E94F: add_psymbol_to_list (psymtab.c:1358)
==8467==    by 0x596634: add_partial_symbol (dwarf2read.c:3794)
==8467==    by 0x596A7A: add_partial_subprogram (dwarf2read.c:3882)

Worth figuring out but it is unrelated to this patch series. So I am 
committing the above patch.