Resurrect Aura Linux accessibility.

Resurrect Aura Linux accessibility.

This is based in part on the ATK code we had as part of the Linux GTK port
that went away, but now rewritten on top of AXPlatformNode. This patch
successfully exposes an accessibility tree for Views and exposes top-level
windows and the the role and name of each View. You can test it using
accerciser. This is just to get the infrastructure in place - we need to
implement more of the interface for it to be useful for accessibility tools.

BUG=463671
Committed: https://crrev.com/f65bad877350b5f7857d29e4b1b0d4d3f5c7cb6d
Cr-Commit-Position: refs/heads/master@{#320383}

[dmazzoni, 2015-03-11 09:00:59]

dmazzoni@chromium.org changed reviewers:
+ plundblad@chromium.org

[dmazzoni, 2015-03-11 09:01:00]

Ready for initial review.

I can't figure out why the clang gn build is failing, I'm specifically
suppressing that warning. Other than that I think it's good.

[Peter Lundblad, 2015-03-11 14:55:00]

lgtm

Just some minor thigns to fix.

[Peter Lundblad, 2015-03-11 14:58:21]

lgtm

Just some minor thigns to fix.

https://codereview.chromium.org/975113002/diff/100001/ui/accessibility/platfo...
File ui/accessibility/platform/atk_util_auralinux.h (right):

https://codereview.chromium.org/975113002/diff/100001/ui/accessibility/platfo...
ui/accessibility/platform/atk_util_auralinux.h:17: public:
nit: indent

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.cc (right):

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:84: if (index < 0
&& index >= GetChildCount())
Should be ||.

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:123:
data_.AddStringAttribute(ui::AX_ATTR_NAME, "Chrome");
Is this a UI string? If so, it should be set from the IDS_SHORT_PRODUCT_NAME
resource I believe so that it depends on the branding.

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.h (right):

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.h:21:
DISALLOW_COPY_AND_ASSIGN(NativeViewAccessibilityAuraLinux);
Missing private: above.

[dmazzoni, 2015-03-11 18:39:53]

https://codereview.chromium.org/975113002/diff/100001/ui/accessibility/platfo...
File ui/accessibility/platform/atk_util_auralinux.h (right):

https://codereview.chromium.org/975113002/diff/100001/ui/accessibility/platfo...
ui/accessibility/platform/atk_util_auralinux.h:17: public:
On 2015/03/11 14:58:21, Peter Lundblad wrote:
> nit: indent

Done.

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.cc (right):

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:84: if (index < 0
&& index >= GetChildCount())
On 2015/03/11 14:58:21, Peter Lundblad wrote:
> Should be ||.

Good catch!

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:123:
data_.AddStringAttribute(ui::AX_ATTR_NAME, "Chrome");
On 2015/03/11 14:58:21, Peter Lundblad wrote:
> Is this a UI string? If so, it should be set from the IDS_SHORT_PRODUCT_NAME
> resource I believe so that it depends on the branding.

Not sure IDS_SHORT_PRODUCT_NAME is available from views, so I plumbed it through
using ViewsDelegate.

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.h (right):

https://codereview.chromium.org/975113002/diff/100001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.h:21:
DISALLOW_COPY_AND_ASSIGN(NativeViewAccessibilityAuraLinux);
On 2015/03/11 14:58:21, Peter Lundblad wrote:
> Missing private: above.

Done.

[dmazzoni, 2015-03-11 22:36:38]

dmazzoni@chromium.org changed reviewers:
+ dpranke@chromium.org, sky@chromium.org

[dmazzoni, 2015-03-11 22:36:39]

+dpranke for build/
+sky for */views/*

[sky, 2015-03-11 23:05:57]

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.cc (right):

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:91: if
(!root_view)
Can this really happen?

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:129:
platform_node_ = ui::AXPlatformNode::Create(this);
move to initializer list?

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:141: };
DISALLOW..

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:147:
AuraLinuxApplication::GetInstance()->RegisterWidget(view->GetWidget());
When is Create() called? If there is no widget at this time is Create() called
again at a later date when there might be a widget?

[Dirk Pranke, 2015-03-11 23:09:15]

build changes lgtm .

[dmazzoni, 2015-03-11 23:38:56]

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
File ui/views/accessibility/native_view_accessibility_auralinux.cc (right):

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:91: if
(!root_view)
On 2015/03/11 23:05:57, sky wrote:
> Can this really happen?

Removed.

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:129:
platform_node_ = ui::AXPlatformNode::Create(this);
On 2015/03/11 23:05:56, sky wrote:
> move to initializer list?

Done.

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:141: };
On 2015/03/11 23:05:56, sky wrote:
> DISALLOW..

Done.

https://codereview.chromium.org/975113002/diff/180001/ui/views/accessibility/...
ui/views/accessibility/native_view_accessibility_auralinux.cc:147:
AuraLinuxApplication::GetInstance()->RegisterWidget(view->GetWidget());
On 2015/03/11 23:05:56, sky wrote:
> When is Create() called? If there is no widget at this time is Create() called
> again at a later date when there might be a widget?

This is typically called as a result of View::NotifyAccessibilityEvent, for
example
when focus changes, and this almost always happens when there's a widget.
If by any chance it's not, the next event fired on a different view within the
same
widget will certainly trigger it.

In manual testing so far this works fine. If it proves to be problematic it
might
be worth looking into exposing a way to enumerate all top-level widgets.

[sky, 2015-03-12 15:28:00]

Ok, LGTM

[dmazzoni, 2015-03-12 21:22:55]

The CQ bit was checked by dmazzoni@chromium.org

[dmazzoni, 2015-03-12 21:22:56]

The patchset sent to the CQ was uploaded after l-g-t-m from
plundblad@chromium.org, dpranke@chromium.org, sky@chromium.org
Link to the patchset: https://codereview.chromium.org/975113002/#ps280001
(title: "Call g_type_init from ax_platform_node_auralinux instead")

[commit-bot: I haz the power, 2015-03-12 21:23:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/975113002/280001

[commit-bot: I haz the power, 2015-03-12 22:00:02]

Committed patchset #15 (id:280001)

[commit-bot: I haz the power, 2015-03-12 22:00:31]

Patchset 15 (id:??) landed as
https://crrev.com/f65bad877350b5f7857d29e4b1b0d4d3f5c7cb6d
Cr-Commit-Position: refs/heads/master@{#320383}

[gunsch, 2015-03-12 22:44:13]

gunsch@chromium.org changed reviewers:
+ gunsch@chromium.org

[gunsch, 2015-03-12 22:44:13]

https://codereview.chromium.org/975113002/diff/280001/ui/views/views.gyp
File ui/views/views.gyp (right):

https://codereview.chromium.org/975113002/diff/280001/ui/views/views.gyp#newc...
ui/views/views.gyp:640: '../../build/linux/system.gyp:atk',
This definition is guarded with "use_ozone==0", but the condition is not.

This breaks Chromecast (trybot in progress), which meets:
OS=="linux"
chromeos==0
use_ozone==1

could you take a look?

[dmazzoni, 2015-03-12 23:14:08]

Sure, I didn't realize we had that combination of flags, I'll fix it to
exclude ozone.

I'm on my way home now and will fix this within a few hours, but if this is
preventing you from working you can revert my change.

On Thu, Mar 12, 2015 at 3:44 PM, <gunsch@chromium.org> wrote:

>
> https://codereview.chromium.org/975113002/diff/280001/ui/views/views.gyp
> File ui/views/views.gyp (right):
>
> https://codereview.chromium.org/975113002/diff/280001/ui/
> views/views.gyp#newcode640
> ui/views/views.gyp:640: '../../build/linux/system.gyp:atk',
> This definition is guarded with "use_ozone==0", but the condition is
> not.
>
> This breaks Chromecast (trybot in progress), which meets:
> OS=="linux"
> chromeos==0
> use_ozone==1
>
> could you take a look?
>
>
> https://codereview.chromium.org/975113002/
>
> To unsubscribe from this group and stop receiving emails from it, send an
> email to chromium-reviews+unsubscribe@chromium.org.
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[alexmos, 2015-03-12 23:16:40]

A revert of this CL (patchset #15 id:280001) has been created in
https://codereview.chromium.org/1002993002/ by alexmos@chromium.org.

The reason for reverting is: Caused compile failure on Linux x64 bot
(http://build.chromium.org/p/chromium.chrome/builders/Google%20Chrome%20Linux%...)

Output snippet:
FAILED: cd ../../chrome; flock -- /tmp/linux_package_lock bash
../out/Release/installer/debian/build.sh -o../out/Release -b ../out/Release -a
x64 -c unstable
dpkg-shlibdeps: warning: binaries to analyze should already be installed in
their package's directory.
2a3
> libatk1.0-0 (>= 1.12.4)

ERROR: Shared library dependencies changed!
If this is intentional, please update:
chrome/installer/linux/debian/expected_deps_ia32
chrome/installer/linux/debian/expected_deps_x64
.