Fix crash in the accessibility code in the shared build. The problem is that the ATL module pointer…

Fix crash in the accessibility code in the shared build. The problem is that the ATL module pointer hasn't been initialized in base.dll when UIATextProvider::CreateTextProvider is called.

To minimize the impact on base.dll, I moved accessibility_misc_utils.* to the ui directory.

BUG=128708
TEST=Build and run chrome shared, access the omni box with a tool that requires accessibility, e.g. the OSK.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=138243

[tommi (sloooow) - chröme, 2012-05-18 11:57:16]

FYI - here's a related change with more info:
https://chromiumcodereview.appspot.com/9958139/

[dmazzoni, 2012-05-18 15:00:27]

The other possibility would be to move accessibility_misc_utils to
ui/base/accessibility, which is actually where Ananta originally had it in his
first draft before I suggested base/win. I didn't think about the fact that base
would then require an ATL module pointer.

Ben, any preference?

https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module.h
File base/win/atl_module.h (right):

https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module...
base/win/atl_module.h:5: #ifndef UI_BASE_WIN_ATL_MODULE_H_
This needs to be updated

https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module...
base/win/atl_module.h:14: namespace ui {
This needs to change from ui to base

[tommi (sloooow) - chröme, 2012-05-18 15:16:36]

On 2012/05/18 15:00:27, Dominic Mazzoni wrote:
> The other possibility would be to move accessibility_misc_utils to
> ui/base/accessibility, which is actually where Ananta originally had it in his
> first draft before I suggested base/win. I didn't think about the fact that
base
> would then require an ATL module pointer.

If it doesn't have to be in base/ then I think it's a good idea to move it out
of there.
However, will we also have to edit content/DEPS to allow including from
ui/base/win?

> 
> Ben, any preference?
> 
>
https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module.h
> File base/win/atl_module.h (right):
> 
>
https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module...
> base/win/atl_module.h:5: #ifndef UI_BASE_WIN_ATL_MODULE_H_
> This needs to be updated
> 
>
https://chromiumcodereview.appspot.com/10399084/diff/7001/base/win/atl_module...
> base/win/atl_module.h:14: namespace ui {
> This needs to change from ui to base

[dmazzoni, 2012-05-18 15:21:35]

> However, will we also have to edit content/DEPS to allow including from
> ui/base/win?

It looks like content/DEPS already allows ui/base, right?

[tommi (sloooow) - chröme, 2012-05-18 16:33:34]

PTAL - I moved the accessibility_misc_utils files to ui/base/win and
updated/reverted the other files accordingly.
I also changed STDMETHODIMP in the accessibility_misc_utils  header file to
STDMETHOD.  The difference between these two macros is that the latter is meant
for declarations (it includes the 'virtual' keyword) whereas the other one is
for the implementation and only includes the return type + calling convention.

[dmazzoni, 2012-05-18 16:41:06]

lgtm

[Ben Goodger (Google), 2012-05-21 17:18:04]

lgtm

[jar (doing other things), 2012-05-21 18:48:09]

LGTM (for removal from base)

[commit-bot: I haz the power, 2012-05-22 07:48:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tommi@chromium.org/10399084/4002

[commit-bot: I haz the power, 2012-05-22 11:36:51]

Try job failure for 10399084-4002 (retry) (retry) on win_rel for steps
"base_unittests, sync_unit_tests".
It's a second try, previously, steps "base_unittests, sync_unit_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...