The patch r172595 was intended for bionic setlocale() which always returns 0. I don't remember I put NULL there initially.
We can simply replace all NULL in r172595 with 0.
I attached the updated patch.
If it is ok, I will commit the patch into google/gcc-4_6 and
google/gcc-4_6-mobile branches.
2011-12-14 H.J. Lu <hongjiu...@intel.com>
Jing Yu <jin...@google.com>
* config/locale/generic/c_locale.h (__convert_from_v): Replace
NULL with 0.
* config/locale/generic/c_locale.cc (__convert_to_v): Likewise
* config/locale/generic/time_members.cc (_M_put): Likewise
Thanks,
Jing
On Wed, Dec 14, 2011 at 10:12 AM, Diego Novillo <dnovi...@google.com> wrote:
> On 11-12-14 13:09 , Paolo Carlini wrote:
>>
>> Hi,
>>
>>> Hi,
>>>
>>> Revision 172595:
>>>
>>> http://gcc.gnu.org/viewcvs?view=revision&revision=172595
>>>
>>> added NULL to config/locale/generic/c_locale.h on
>>> google/gcc-4_6-branch. But<cstddef> isn't included, which lead to
>>> build failure since NULL isn't defined. This patch
>>> includes<cstddef>. OK for google/gcc-4_6-branch?
>>
>>
>> If you are interested in my personal opinion, I have of course
>> nothing to do with this branch, I think this code should not simply
>> use NULL. After all, its C++, right? ;)
>
>
> Seems reasonable. Jing, 172595 was a patch from you. Would it be the same
> if you just used 0 instead of NULL?
>
>
> Diego.
patchnull.diff
Description: Binary data
