Bug 148967 – AX: ARIA 1.1 implement aria-colcount, aria-colindex, aria-colspan, aria-rowcount, aria-rowindex and aria-rowspan

• Component: Accessibility
• Status: RESOLVED FIXED

[Nan Wang]

Added support for ARIA 1.1 table/grid related attributes.

aria-colcount
http://www.w3.org/TR/wai-aria-1.1/#aria-colcount
Defines an element's column index or position with respect to the total number of columns within a table, grid, or treegrid. 

aria-colindex
http://www.w3.org/TR/wai-aria-1.1/#aria-colindex
Defines an element's column index or position with respect to the total number of columns within a table, grid, or treegrid. S

aria-colspan
http://www.w3.org/TR/wai-aria-1.1/#aria-colspan
Defines the number of columns spanned by a cell or gridcell within a table, grid, or tree grid.

aria-rowcount
http://www.w3.org/TR/wai-aria-1.1/#aria-rowcount
Defines the total number of rows in a table, grid, or tree grid.

aria-rowindex
http://www.w3.org/TR/wai-aria-1.1/#aria-rowindex
Defines an element's row index or position with respect to the total number of rows within a table, grid, or tree grid.

aria-rowspan
http://www.w3.org/TR/wai-aria-1.1/#aria-rowspan
Defines the number of rows spanned by a cell or gridcell within a table, grid, or treegrid.

[Radar WebKit Bug Importer]

<rdar://problem/22612833>

[Nan Wang]

Created attachment 262787 [details]
patch

I don't want to mess up with the native table attributes/properties (like cell index ranges, table column/row counts, etc.), so added several new attributes for mac to query for these new table/grid related values. 

On mac, maybe VoiceOver can check these new attributes first then decide what to speak.

[Nan Wang]

Created attachment 262789 [details]
patch

Fixed iOS build failure.

[chris fleizach]

Comment on attachment 262789 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=262789&action=review

whats the plan for exposing this on iOS?

> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:96
> +    unsigned rowSpan = ariaRowSpan();

how do we distinguish 0 from no value?

> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:101
> +        rowSpan = 1;

can you put this blob into a separate method

> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:145
> +}

i know that column/rowIndex are used by VO to get data out of the table sometimes. will this mess things up?

> Source/WebCore/accessibility/AccessibilityTable.cpp:695
> +    int rowCountInt = rowCountValue.toInt();

if there's no aria_rowCount set, won't toInt() == 0? which seems like the wrong default value right

> Source/WebCore/accessibility/AccessibilityTableRow.cpp:152
> +    for (unsigned i = 0; i < rowChildren.size(); i++) {

are we able to use the better iterators for accessibility children

for (const auto& cell in rowChildren)

[Nan Wang]

Comment on attachment 262789 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=262789&action=review

For iOS I'm thinking of adding new methods to expose those values. I might open a new bug for iOS and tests. But seems iOS is not speaking table row/column count or indexes now.

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:96
>> +    unsigned rowSpan = ariaRowSpan();
> 
> how do we distinguish 0 from no value?

If no value, the function will return 1 as default.

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:101
>> +        rowSpan = 1;
> 
> can you put this blob into a separate method

Ok

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:145
>> +}
> 
> i know that column/rowIndex are used by VO to get data out of the table sometimes. will this mess things up?

This is only for non-native tables, so I think the colspan/rowspan attributes won't be affected from this code. If author didn't set aria-colspan/aria-rowspan, the range.second will be 1 as before.

>> Source/WebCore/accessibility/AccessibilityTable.cpp:695
>> +    int rowCountInt = rowCountValue.toInt();
> 
> if there's no aria_rowCount set, won't toInt() == 0? which seems like the wrong default value right

My idea is to return the value only if the aria-rowcount is greater than the rendered table row count. Otherwise just return -1, and VO doesn't have to worry about speaking it.
The spec says: authors must set the value of aria-rowcount to -1 to indicate that the value should not be calculated by the user agent. Not sure if we should do something special about it.

>> Source/WebCore/accessibility/AccessibilityTableRow.cpp:152
>> +    for (unsigned i = 0; i < rowChildren.size(); i++) {
> 
> are we able to use the better iterators for accessibility children
> 
> for (const auto& cell in rowChildren)

Ok

[Nan Wang]

Comment on attachment 262789 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=262789&action=review

For iOS I'm thinking of adding new methods to expose those values. I might open a new bug for iOS and tests. But seems iOS is not speaking table row/column count or indexes now.

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:96
>> +    unsigned rowSpan = ariaRowSpan();
> 
> how do we distinguish 0 from no value?

If no value, the function will return 1 as default.

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:101
>> +        rowSpan = 1;
> 
> can you put this blob into a separate method

Ok

>> Source/WebCore/accessibility/AccessibilityARIAGridCell.cpp:145
>> +}
> 
> i know that column/rowIndex are used by VO to get data out of the table sometimes. will this mess things up?

This is only for non-native tables, so I think the colspan/rowspan attributes won't be affected from this code. If author didn't set aria-colspan/aria-rowspan, the range.second will be 1 as before.

>> Source/WebCore/accessibility/AccessibilityTable.cpp:695
>> +    int rowCountInt = rowCountValue.toInt();
> 
> if there's no aria_rowCount set, won't toInt() == 0? which seems like the wrong default value right

My idea is to return the value only if the aria-rowcount is greater than the rendered table row count. Otherwise just return -1, and VO doesn't have to worry about speaking it.
The spec says: authors must set the value of aria-rowcount to -1 to indicate that the value should not be calculated by the user agent. Not sure if we should do something special about it.

>> Source/WebCore/accessibility/AccessibilityTableRow.cpp:152
>> +    for (unsigned i = 0; i < rowChildren.size(); i++) {
> 
> are we able to use the better iterators for accessibility children
> 
> for (const auto& cell in rowChildren)

Ok

[Nan Wang]

Created attachment 262797 [details]
patch

[chris fleizach]

Comment on attachment 262797 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=262797&action=review

> Source/WebCore/accessibility/AccessibilityTableCell.cpp:376
> +    const AtomicString& colIndexValue = getAttribute(aria_colindexAttr);

should these methods verify that the table cell has a role="grid cell"

[chris fleizach]

Comment on attachment 262797 [details]
patch

ok

[WebKit Commit Bot]

Comment on attachment 262797 [details]
patch

Clearing flags on attachment: 262797

Committed r190833: <http://trac.webkit.org/changeset/190833>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[James Craig]

Comment on attachment 262797 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=262797&action=review

>> Source/WebCore/accessibility/AccessibilityTableCell.cpp:376
>> +    const AtomicString& colIndexValue = getAttribute(aria_colindexAttr);
> 
> should these methods verify that the table cell has a role="grid cell"

I don't think so. These could work with "cell", "gridcell" (when appropriate), or the implicit role from the th/td.

[Nan Wang]

Created attachment 263325 [details]
example html

Added an example html for testing.

[James Craig]

*** Bug 163644 has been marked as a duplicate of this bug. ***