Fix a bunch of clang warnings/errors.

Fix a bunch of clang warnings/errors.

BUG=None
TEST=None

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=60109

[Nico, 2010-09-21 00:19:42]

Mostly "unused function", but also some other stuff.

[Mark Mentovai, 2010-09-21 18:40:01]

LGTM. Windows try failure doesn’t look relevant.

http://codereview.chromium.org/3473006/diff/1/5
File base/scoped_nsobject.h (right):

http://codereview.chromium.org/3473006/diff/1/5#newcode86
base/scoped_nsobject.h:86: void swap(scoped_nsobject<C>& p1, scoped_nsobject<C>&
p2) {
What uses swap as a free guy?

http://codereview.chromium.org/3473006/diff/1/6
File base/third_party/dmg_fp/dtoa.cc (right):

http://codereview.chromium.org/3473006/diff/1/6#newcode1562
base/third_party/dmg_fp/dtoa.cc:1562: if (!hexdig[(int)'0'])
Edit README.chromium and provide a patch or modify an existing patch in this
directory.

[Evan Martin, 2010-09-21 18:52:20]

I think we should disable the unused function warning for now.

http://codereview.chromium.org/3473006/diff/1/2
File app/l10n_util.cc (right):

http://codereview.chromium.org/3473006/diff/1/2#newcode230
app/l10n_util.cc:230: #if !defined(OS_MACOSX)
I think we disable the unused function warnings already with gcc.  We have way
too much code to need to worry about which functions are used on which
platforms.  (I recall there being random functions in skia that aren't used for
example.)

http://codereview.chromium.org/3473006/diff/1/6
File base/third_party/dmg_fp/dtoa.cc (right):

http://codereview.chromium.org/3473006/diff/1/6#newcode1562
base/third_party/dmg_fp/dtoa.cc:1562: if (!hexdig[(int)'0'])
On 2010/09/21 18:40:01, Mark Mentovai wrote:
> Edit README.chromium and provide a patch or modify an existing patch in this
> directory.

I intentionally didn't make this change when I did the original landing.  We
should probably just disable the warning until we understand why it's even
warning about this.

http://codereview.chromium.org/3473006/diff/1/14
File chrome/browser/spellcheck_host.cc (right):

http://codereview.chromium.org/3473006/diff/1/14#newcode34
chrome/browser/spellcheck_host.cc:34: #if defined(OS_WIN)
this one too

[Nico, 2010-09-21 20:26:21]

On Tue, Sep 21, 2010 at 11:52 AM,  <evan@chromium.org> wrote:
> I think we should disable the unused function warning for now.
>
>
> http://codereview.chromium.org/3473006/diff/1/2
> File app/l10n_util.cc (right):
>
> http://codereview.chromium.org/3473006/diff/1/2#newcode230
> app/l10n_util.cc:230: #if !defined(OS_MACOSX)
> I think we disable the unused function warnings already with gcc.  We
> have way too much code to need to worry about which functions are used
> on which platforms.  (I recall there being random functions in skia that
> aren't used for example.)

It successfully compiles with the warning for me by now. In the
process I removed an unused function in googleurl and there's quite a
bit of red in this patch because of it. Seems like having the warning
is a good thing?

>
> http://codereview.chromium.org/3473006/diff/1/6
> File base/third_party/dmg_fp/dtoa.cc (right):
>
> http://codereview.chromium.org/3473006/diff/1/6#newcode1562
> base/third_party/dmg_fp/dtoa.cc:1562: if (!hexdig[(int)'0'])
> On 2010/09/21 18:40:01, Mark Mentovai wrote:
>>
>> Edit README.chromium and provide a patch or modify an existing patch
>
> in this
>>
>> directory.
>
> I intentionally didn't make this change when I did the original landing.
>  We should probably just disable the warning until we understand why
> it's even warning about this.

Ok.

>
> http://codereview.chromium.org/3473006/diff/1/14
> File chrome/browser/spellcheck_host.cc (right):
>
> http://codereview.chromium.org/3473006/diff/1/14#newcode34
> chrome/browser/spellcheck_host.cc:34: #if defined(OS_WIN)
> this one too
>
> http://codereview.chromium.org/3473006/show
>

[Evan Martin, 2010-09-21 20:51:31]

On 2010/09/21 20:26:21, Nico wrote:
> It successfully compiles with the warning for me by now. In the
> process I removed an unused function in googleurl and there's quite a
> bit of red in this patch because of it. Seems like having the warning
> is a good thing?

Wow, these are the only unused functions?  I guess it's ok then