Replace usage of MFC wrapper classes with their win32 base types to reduce de...

Replace usage of MFC wrapper classes with their win32 base types to reduce dependency on ATL/MFC.

BUG=5027
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=25298

[Peter Kasting, 2009-09-02 23:42:15]

OK, but can be even better by using some of our non-ATL scoping objects.

http://codereview.chromium.org/171129/diff/4003/3002
File app/win_util.cc (right):

http://codereview.chromium.org/171129/diff/4003/3002#newcode13
Line 13: #include <algorithm>
Nit: Why is this needed?

http://codereview.chromium.org/171129/diff/4003/3002#newcode694
Line 694: HDC mem_hdc = CreateCompatibleDC(window_hdc);
Nit: Use ScopedHDC (see scoped_handle_win.h)

http://codereview.chromium.org/171129/diff/4003/3002#newcode705
Line 705: HBITMAP bitmap = CreateDIBSection(mem_hdc,
Nit: Use ScopedGDIObject (see scoped_handle_win.h)

[James Hawkins, 2009-09-03 00:10:44]

http://codereview.chromium.org/171129/diff/4003/3002
File app/win_util.cc (right):

http://codereview.chromium.org/171129/diff/4003/3002#newcode13
Line 13: #include <algorithm>
On 2009/09/02 23:42:16, Peter Kasting wrote:
> Nit: Why is this needed?

std::max is used later on in the file.  This is from the include-what-you-use
lint warning.

http://codereview.chromium.org/171129/diff/4003/3002#newcode694
Line 694: HDC mem_hdc = CreateCompatibleDC(window_hdc);
On 2009/09/02 23:42:16, Peter Kasting wrote:
> Nit: Use ScopedHDC (see scoped_handle_win.h)

Done.

http://codereview.chromium.org/171129/diff/4003/3002#newcode705
Line 705: HBITMAP bitmap = CreateDIBSection(mem_hdc,
On 2009/09/02 23:42:16, Peter Kasting wrote:
> Nit: Use ScopedGDIObject (see scoped_handle_win.h)

Done.  Used ScopedBitmap (specialization of ScopedGDIObject).

[Peter Kasting, 2009-09-03 04:24:48]

LGTM

http://codereview.chromium.org/171129/diff/3003/8001
File app/win_util.cc (right):

http://codereview.chromium.org/171129/diff/3003/8001#newcode742
Line 742: ReleaseDC(window_handle, window_hdc);
BTW, we don't have any problems due to destruction order, right?  Windows won't
care that window_hdc gets released before mem_hdc or something?  I don't think
so, just being paranoid.

[James Hawkins, 2009-09-03 04:51:31]

On 2009/09/03 04:24:48, Peter Kasting wrote:
> LGTM
> 
> http://codereview.chromium.org/171129/diff/3003/8001
> File app/win_util.cc (right):
> 
> http://codereview.chromium.org/171129/diff/3003/8001#newcode742
> Line 742: ReleaseDC(window_handle, window_hdc);
> BTW, we don't have any problems due to destruction order, right?  Windows
won't
> care that window_hdc gets released before mem_hdc or something?  I don't think
> so, just being paranoid.

It's safe because mem_hdc is a separate copy and releasing window_hdc won't
destroy the window DC anyway.