Coverity: Initialize member variables.

Coverity: Initialize member variables.

CID_COUNT=13
CID=102812,102838,102863,102867,102875,102891,102906,102924,102947,102964,
    103070,103246,103259
BUG=none
TEST=none
R=groby

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=120913

[Sergey Ulanov, 2012-02-05 09:48:53]

remoting/* LGTM

[scherkus (not reviewing), 2012-02-06 17:40:41]

LGTM on content/common/gpu/media/dxva_video_decode_accelerator.cc

[dmichael (off chromium), 2012-02-06 18:38:07]

ppapi/* LGTM

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
File ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc
(right):

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc:15:
memset(&instance_1_1_, 0, sizeof(instance_1_1_));
doesn't it suffice to just put instance_1_1_() in the initializer list?

[groby-ooo-7-16, 2012-02-06 18:39:48]

LGTM - modulo resolving the perf question for TraceEvents

http://codereview.chromium.org/9307114/diff/6001/base/debug/trace_event_impl.cc
File base/debug/trace_event_impl.cc (right):

http://codereview.chromium.org/9307114/diff/6001/base/debug/trace_event_impl....
base/debug/trace_event_impl.cc:148: memset(arg_names_, 0, sizeof(arg_names_));
Technically, the old way was OK - those are both strings. I wouldn't usually
mind the memset, but TraceEvent is performance critical IIRC :)

[James Hawkins, 2012-02-07 21:09:20]

https://chromiumcodereview.appspot.com/9307114/diff/6001/base/debug/trace_eve...
File base/debug/trace_event_impl.cc (right):

https://chromiumcodereview.appspot.com/9307114/diff/6001/base/debug/trace_eve...
base/debug/trace_event_impl.cc:148: memset(arg_names_, 0, sizeof(arg_names_));
On 2012/02/06 18:39:48, groby wrote:
> Technically, the old way was OK - those are both strings. I wouldn't usually
> mind the memset, but TraceEvent is performance critical IIRC :)

You're right, and Coverity wasn't even complaining about |arg_names_| anyway.

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
File ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc
(right):

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc:15:
memset(&instance_1_1_, 0, sizeof(instance_1_1_));
On 2012/02/06 18:38:07, dmichael wrote:
> doesn't it suffice to just put instance_1_1_() in the initializer list?

The default constructor doesn't initialize the POD types (in this case, the
function pointers).

[dmichael (off chromium), 2012-02-07 21:30:19]

lgtm

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
File ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc
(right):

https://chromiumcodereview.appspot.com/9307114/diff/6001/ppapi/native_client/...
ppapi/native_client/src/shared/ppapi_proxy/ppp_instance_combined.cc:15:
memset(&instance_1_1_, 0, sizeof(instance_1_1_));
On 2012/02/07 21:09:21, James Hawkins wrote:
> On 2012/02/06 18:38:07, dmichael wrote:
> > doesn't it suffice to just put instance_1_1_() in the initializer list?
> 
> The default constructor doesn't initialize the POD types (in this case, the
> function pointers).
instance_1_1_ is a POD struct. Adding instance_1_1_() to the initializer list
"value-initializes" instance_1_1_, which means the pointers are all set to NULL.
See, e.g.:
http://stackoverflow.com/questions/3803153/what-are-primitive-types-default-i...

That said, not everyone knows that, so memset may be clearer. I'm fine either
way.