Build fix for [-Werror=maybe-uninitialized].

Build fix for [-Werror=maybe-uninitialized].

gcc v4.7.3 failed to build due to uninitialized variables.

BUG=http://code.google.com/p/skia/issues/detail?id=1381

Committed: http://code.google.com/p/skia/source/detail?r=9887

[dshwang, 2013-06-27 15:22:54]

ping

[bungeman-skia, 2013-06-27 17:21:14]

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h
File src/core/SkPathRef.h (right):

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h#newcode484
src/core/SkPathRef.h:484: pCnt = 0;
It may be better to put a default here to set pCnt to 0 (to avoid the warning)
and SkASSERT.

reed: Does this need to be updated for conics?

[bungeman-skia, 2013-06-27 17:29:23]

I have opened https://code.google.com/p/skia/issues/detail?id=1381 to track
having a bot which runs the latest (or at least a newer) version of gcc.

[reed1, 2013-06-27 17:34:14]

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h
File src/core/SkPathRef.h (right):

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h#newcode484
src/core/SkPathRef.h:484: pCnt = 0;
On 2013/06/27 17:21:14, bungeman1 wrote:
> It may be better to put a default here to set pCnt to 0 (to avoid the warning)
> and SkASSERT.
> 
> reed: Does this need to be updated for conics?

Other than line 473 ?

[dshwang, 2013-06-27 17:54:33]

Thx for review.

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h
File src/core/SkPathRef.h (right):

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h#newcode484
src/core/SkPathRef.h:484: pCnt = 0;
On 2013/06/27 17:21:14, bungeman1 wrote:
> It may be better to put a default here to set pCnt to 0 (to avoid the warning)
> and SkASSERT.
> 
> reed: Does this need to be updated for conics?

Do you mean kClose_Verb must be not reached also?

[bungeman-skia, 2013-06-27 19:06:03]

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h
File src/core/SkPathRef.h (right):

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h#newcode484
src/core/SkPathRef.h:484: pCnt = 0;
On 2013/06/27 17:54:33, dshwang wrote:
> On 2013/06/27 17:21:14, bungeman1 wrote:
> > It may be better to put a default here to set pCnt to 0 (to avoid the
warning)
> > and SkASSERT.
> > 
> > reed: Does this need to be updated for conics?
> 
> Do you mean kClose_Verb must be not reached also?

No, I suppose the change as stated is probably fine, in a release build I'm sure
all modern compilers will assign pCnt once or decide that it's faster to just do
the two independent writes. My comment just shows that I like the structure
which makes this look like a simple assignment to a 'const int pCnt'.

It would be nice if gcc noted that we're covering all of the enum values here,
but it apparently doesn't care about that. The docs at
http://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html recommend using a default
with an assert.

In any event this is really reed's code.

https://codereview.chromium.org/17572010/diff/1/src/core/SkPathRef.h#newcode484
src/core/SkPathRef.h:484: pCnt = 0;
On 2013/06/27 17:34:14, reed1 wrote:
> On 2013/06/27 17:21:14, bungeman1 wrote:
> > It may be better to put a default here to set pCnt to 0 (to avoid the
warning)
> > and SkASSERT.
> > 
> > reed: Does this need to be updated for conics?
> 
> Other than line 473 ?

Oops, missed that fall through when scanning. My OCD is acting up now because
Conic/Quad and Close/Done are out of declared order here.

[dshwang, 2013-07-03 07:50:26]

Thx for detail explanation.

https://codereview.chromium.org/17572010/diff/13001/src/core/SkPathRef.h
File src/core/SkPathRef.h (right):

https://codereview.chromium.org/17572010/diff/13001/src/core/SkPathRef.h#newc...
src/core/SkPathRef.h:489: pCnt = 0;
How about this change.
now case statements are sorted by enum order.
pCnt is initialized only one time.

[bungeman-skia, 2013-07-03 14:26:46]

lgtm

[commit-bot: I haz the power, 2013-07-03 15:02:20]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/dongseong.hwang@intel.com/17572010/13001

[commit-bot: I haz the power, 2013-07-03 15:09:32]

Change committed as 9887