Fixing some warnings on Linux

Fixing some warnings on Linux

There were only a few warnings left, so I fixed them and enabled the unused variable warning by removing the "-Wno-unused" flag. Only the -Wno-unused-parameter remains for now (could be removed later).
Committed: https://code.google.com/p/skia/source/detail?r=8030

[sugoi, 2013-03-05 21:21:27]

https://codereview.chromium.org/12480002/diff/1/src/gpu/gr_unittests.cpp
File src/gpu/gr_unittests.cpp (right):

https://codereview.chromium.org/12480002/diff/1/src/gpu/gr_unittests.cpp#newc...
src/gpu/gr_unittests.cpp:39: GrAssert(index < 0);
This code does nothing in RELEASE since all these GrAssert calls are DEBUG only,
so it can all get optimized out.

[sugoi, 2013-03-05 21:39:35]

https://codereview.chromium.org/12480002/diff/1/gyp/utils.gyp
File gyp/utils.gyp (right):

https://codereview.chromium.org/12480002/diff/1/gyp/utils.gyp#newcode252
gyp/utils.gyp:252: ['skia_os in ["android", "linux", "freebsd", "openbsd",
"solaris", "nacl"]', {
Oops, "android" shouldn't be in this list. Removed it.

[sugoi, 2013-03-05 21:51:16]

I sent the patch to all non mac, non windows trybots. Hopefully, I'll get all
answers by tomorrow.

[bsalomon, 2013-03-05 21:57:54]

On 2013/03/05 21:51:16, sugoi wrote:
> I sent the patch to all non mac, non windows trybots. Hopefully, I'll get all
> answers by tomorrow.

lgtm

[sugoi, 2013-03-06 16:14:59]

I'm going to ignore the original lgtm, as this cl has become a lot larger now
that I fixed it for the buildbots. I fixed warnings for :
- Animator
- Sample code
- Views
- 32 bit specific warnings
- Non GPU specific warnings

https://codereview.chromium.org/12480002/diff/8009/include/views/SkWidget.h
File include/views/SkWidget.h (right):

https://codereview.chromium.org/12480002/diff/8009/include/views/SkWidget.h#n...
include/views/SkWidget.h:106: virtual Click* onFindClickHandler(SkScalar x,
SkScalar y, unsigned modi);
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleLines.cpp
File samplecode/SampleLines.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleLines.cpp...
samplecode/SampleLines.cpp:101: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleOvalTest.cpp
File samplecode/SampleOvalTest.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleOvalTest....
samplecode/SampleOvalTest.cpp:104: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SamplePathClip.cpp
File samplecode/SamplePathClip.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SamplePathClip....
samplecode/SamplePathClip.cpp:73: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleSlides.cpp
File samplecode/SampleSlides.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleSlides.cp...
samplecode/SampleSlides.cpp:720: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleTextAlpha...
File samplecode/SampleTextAlpha.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleTextAlpha...
samplecode/SampleTextAlpha.cpp:97: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleUnitMappe...
File samplecode/SampleUnitMapper.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleUnitMappe...
samplecode/SampleUnitMapper.cpp:144: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleVertices.cpp
File samplecode/SampleVertices.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/samplecode/SampleVertices....
samplecode/SampleVertices.cpp:123: virtual SkView::Click*
onFindClickHandler(SkScalar x, SkScalar y, unsigned) {
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/8009/src/animator/SkAnimator.cpp
File src/animator/SkAnimator.cpp (right):

https://codereview.chromium.org/12480002/diff/8009/src/animator/SkAnimator.cp...
src/animator/SkAnimator.cpp:412: SkDEBUGCODE(bool success =) evt.findPtr("anim",
(void**) &eventState.fDisplayable);
variable success is only used in debug code

https://codereview.chromium.org/12480002/diff/21004/src/animator/SkDrawExtraP...
File src/animator/SkDrawExtraPathEffect.cpp (left):

https://codereview.chromium.org/12480002/diff/21004/src/animator/SkDrawExtraP...
src/animator/SkDrawExtraPathEffect.cpp:121: SkDrawMatrix* matrix;
matrix is assigned, but unused, so it was deleted

https://codereview.chromium.org/12480002/diff/21004/src/animator/SkDrawExtraP...
File src/animator/SkDrawExtraPathEffect.cpp (right):

https://codereview.chromium.org/12480002/diff/21004/src/animator/SkDrawExtraP...
src/animator/SkDrawExtraPathEffect.cpp:231: virtual void begin(const SkIRect&
uvBounds, SkPath*) const {
These virtual functions are const in the base class, so I made them const here
too. In order to keep the original logic without doing anything illegal, I put
the data into a data struct so that the data may be modified while keeping these
functions const.

https://codereview.chromium.org/12480002/diff/21004/src/pdf/SkPDFCatalog.cpp
File src/pdf/SkPDFCatalog.cpp (right):

https://codereview.chromium.org/12480002/diff/21004/src/pdf/SkPDFCatalog.cpp#...
src/pdf/SkPDFCatalog.cpp:132: SkASSERT(fCatalog[i].fFileOffset < 2147483647);
On 32 bits platforms, off_t is a 32 bits signed integer, so we have to test
against a slightly smaller value to make sure it's always valid, otherwise, the
test will always be true on these platforms. I'm assuming 9999999999 was an
arbitrarily large number anyway.

https://codereview.chromium.org/12480002/diff/21004/src/ports/SkFontHost_font...
File src/ports/SkFontHost_fontconfig.cpp (left):

https://codereview.chromium.org/12480002/diff/21004/src/ports/SkFontHost_font...
src/ports/SkFontHost_fontconfig.cpp:252: int index;
unused variable

https://codereview.chromium.org/12480002/diff/21004/src/views/SkWidgets.cpp
File src/views/SkWidgets.cpp (right):

https://codereview.chromium.org/12480002/diff/21004/src/views/SkWidgets.cpp#n...
src/views/SkWidgets.cpp:312: SkView::Click*
SkPushButtonWidget::onFindClickHandler(SkScalar x, SkScalar y, unsigned modi)
virtual function's signature did not fit base class' signature

https://codereview.chromium.org/12480002/diff/21004/src/views/animated/SkScro...
File src/views/animated/SkScrollBarView.cpp (right):

https://codereview.chromium.org/12480002/diff/21004/src/views/animated/SkScro...
src/views/animated/SkScrollBarView.cpp:125: //    int hideBar = 0;
The line using this variable is commented out.

https://codereview.chromium.org/12480002/diff/21004/src/views/unix/SkOSWindow...
File src/views/unix/SkOSWindow_Unix.cpp (left):

https://codereview.chromium.org/12480002/diff/21004/src/views/unix/SkOSWindow...
src/views/unix/SkOSWindow_Unix.cpp:158: long event_mask = NoEventMask;
unused variable

[sugoi, 2013-03-07 15:16:37]

Ok, so for treating warnings as errors, I chickened out for now. I changed this
cl to just fix some warnings and enable the unused variable warning, which are
almost all removed by this cl. There's 1 left, which I'll fix later.

[bsalomon, 2013-03-07 15:20:58]

On 2013/03/07 15:16:37, sugoi wrote:
> Ok, so for treating warnings as errors, I chickened out for now. I changed
this
> cl to just fix some warnings and enable the unused variable warning, which are
> almost all removed by this cl. There's 1 left, which I'll fix later.

Mostly lgtm but two things:

1) Someone should confirm the PDF 999...9LL change

2) Everytwhere you added the unsigned to the click handler function can you also
add SK_OVERRIDE?

[sugoi, 2013-03-07 15:34:59]

On 2013/03/07 15:20:58, bsalomon wrote:
> 1) Someone should confirm the PDF 999...9LL change
I can wait for Mike's input, but I think the assert as is it is wrong. If this
off_t value can exceed the 32 bits signed max value, then this means it will
fail on 32 bits platforms, so I though that this probably shouldn't be allowed.
Also, a file offset larger than 2 Gb in a PDF file is probably unlikely to
happen.

> 2) Everytwhere you added the unsigned to the click handler function can you
also add SK_OVERRIDE?
Will do, no problem. The only weird thing is that other virtual functions around
these do not have the SK_OVERRIDE tag, so that does look a bit odd.

[bsalomon, 2013-03-07 15:47:33]

On 2013/03/07 15:34:59, sugoi wrote:
> On 2013/03/07 15:20:58, bsalomon wrote:

> > 2) Everytwhere you added the unsigned to the click handler function can you
> also add SK_OVERRIDE?
> Will do, no problem. The only weird thing is that other virtual functions
around
> these do not have the SK_OVERRIDE tag, so that does look a bit odd.

SK_OVERRIDE didn't exist when these headers were written. I think it's better to
have it on some overrides than none but feel free to add it to the others if you
like. :)

[sugoi, 2013-03-07 15:55:49]

On 2013/03/07 15:47:33, bsalomon wrote:
> SK_OVERRIDE didn't exist when these headers were written. I think it's better
to have it on some overrides than none but feel free to add it to the others if
you like. :)

Maybe some other time :) I looked it up and the PDF 999...9LL change was
originally written by vandebo, so I just sent him an e-mail to ask him if it was
really important for this value to be around 10 Gb rather then 2 Gb. If he
agrees that 2 Gb is ok, I'll commit the code.

[sugoi, 2013-03-07 20:52:02]

I'm having some trouble using the buildbots for some reason, so I just removed
SkPDFCatalog.cpp from the cl. I'll commit the rest of it now and deal with this
file later.

[sugoi1, 2013-03-07 20:53:09]

Committed patchset #15 manually as r8030 (presubmit successful).