On Fri, Dec 13, 2013 at 4:36 PM, Luca Barbato <lu_z...@gentoo.org> wrote:
> On 13/12/13 16:15, Vittorio Giovara wrote:
>> On Fri, Dec 13, 2013 at 4:13 PM, Luca Barbato <lu_z...@gentoo.org> wrote:
>>> On 13/12/13 16:12, Vittorio Giovara wrote:
>>>> On Fri, Dec 13, 2013 at 12:28 AM, Guillaume Martres <smar...@ubuntu.com>
>>>> wrote:
>>>>> From: Michael Niedermayer <michae...@gmx.at>
>>>>>
>>>>> ---
>>>>> libavcodec/hevc_sei.c | 5 ++++-
>>>>> 1 file changed, 4 insertions(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/libavcodec/hevc_sei.c b/libavcodec/hevc_sei.c
>>>>> index eed7b33..88b4744 100644
>>>>> --- a/libavcodec/hevc_sei.c
>>>>> +++ b/libavcodec/hevc_sei.c
>>>>> @@ -49,7 +49,10 @@ static void
>>>>> decode_nal_sei_decoded_picture_hash(HEVCContext *s)
>>>>> static void decode_nal_sei_frame_packing_arrangement(HEVCContext *s)
>>>>> {
>>>>> GetBitContext *gb = &s->HEVClc.gb;
>>>>> - int cancel, type, quincunx, content;
>>>>> + int cancel;
>>>>> + int quincunx = 0;
>>>>> + int content = -1;
>>>>> + int type = -1;
>>>>
>>>> Actually these variables are not used at all when unintialized, as
>>>> they are all guarded by proper "frame_packing_present_flag" checks.
>>>> What this patch fixes is just a few warnings from clang, I'll amend
>>>> the commit message to reflect this.
>>>
>>> wait, why clang is issuing warnings? the static analyzer should tell us
>>> more.
>>>
>>
>> It's the "sometimes-uninitialized" warning that we discussed some time ago.
>> Vittorio
>
> maybe-uninitialized should be suppressed by the default cflags.
clang doesn't come with "maybe-uninitialized", it has
"sometimes-uninitialized" which is not the same thing.
Vittorio
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
