Assigned ax role for table headers.

Assigned ax role for table headers.

This patch makes <th> node have column header or row header.
The basic rule is as follows.
If header is preceded by header cells then it's a column header,
if it is preceded by cells then it's a row header.
If header is followed by header cells then it's a column header,
if it is followed by cells then it's a row header.

BUG=435081
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=187901

[je_julie(Not used), 2015-01-05 07:16:47]

je_julie.kim@samsung.com changed reviewers:
+ dmazzoni@chromium.org

[je_julie(Not used), 2015-01-05 07:16:47]

Hi Dominic,
This patch makes <th> node have column header or row header according to sibling
node.
The basic rule is from behavior of Firefox.
Because it changes role from cell to header for <th>, we have to disable several
TCs from content_browsertests.
Here is the patch, crrev.com/823293003, for disabling those TCs.

PTAL.

[dmazzoni, 2015-01-06 00:04:58]

lgtm

The patch looks great.

One potential issue that worries me is what happens if the table gets modified
dynamically. If a cell gets a new sibling, it may have to recompute its role.
Could you file a bug for that?

[je_julie(Not used), 2015-01-06 00:22:33]

On 2015/01/06 00:04:58, dmazzoni wrote:
> lgtm
> 
> The patch looks great.
> 
> One potential issue that worries me is what happens if the table gets modified
> dynamically. If a cell gets a new sibling, it may have to recompute its role.
> Could you file a bug for that?

Sure. I'll make a test case for that and file a bug.
Thanks.

[je_julie(Not used), 2015-01-06 06:33:10]

The CQ bit was checked by je_julie.kim@samsung.com

[commit-bot: I haz the power, 2015-01-06 06:33:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/816423003/1

[commit-bot: I haz the power, 2015-01-06 07:49:34]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=187901