Move core getRecord logic from view -> core getRecord was passing through to this.body.getRecord(args) but that required the table to be rendered. Accessing the records by id or clientId shouldn't depend on the rendered state.

Don't use clientId <tr id="HERE">. Fixes #2531889

+syncUI renders t(head/body/foot) content

Clean up module vs class descriptions, +@since 3.5.0 Moved the lengthy text from the module description to the class description so the module page would look cleaner. Added note for each class extension to point readers to the DataTable class for docs on the APIs contributed by this class, since they all have @for DataTable.

Preserve first column assignment to map if dups If more than one column have the same key, getColumn({key}) would return the last one, and there was no way to get the first.

Fix set('data', modelListInstance)

Add back compat for field and id column configs core.js added the back compat. The others share an update of changing references of col._yuid to col.id.

Don't set table width if width attribute is unset.

_uiSetWidth needs to account for border width This unfortunately breaks percentage widths for non-scrolling tables, since the browser resize is not tracked in core. The initial paint should be fine, but resizing the browser won't resize the table. I need to move that logic from datatable-scroll to core, but it'll have to wait for a future sprint :(

Last changes to get datatable-scroll working * Added cellspacing="0" to core and fixed header template and removed *border-collapse: collapse hack for IE * Double wrapped the Y scrolling nodes. New <div> has the border and contains the fixed headers and the overflow <div>. This obviated some dim calculations to work around border widths for placement of the fixed header. Also allowed an overflow-x: hidden rule to hide any column width rounding issues resulting in the fixed headers being 1px too wide and overflowing the previous yScrollNode border. * Use table.replace(wrapper).appendTo(wrapper) instead of table.wrap(wrapper) because IE (I forget which version--6 or 7, I think) did all sorts of stupid things, like extracting the last table cell and wrapping the rest of the table inside it. It's probably a bug that needs addressing in Node.prototype.wrap, but this solution also works. I do believe this will be the shipped version of datatable-scroll. Hurrah!

Add support for percentage widths + x scrolling

Remove _parent property from moved columns

Forgot a return statement. Hurray for tests.

Move col._parent assignment to Core Needed for getColumn.

Functional support for recordset, columnset attrs For back compat. Supports assignment only (construction time or via set(...)). get('recordset/columnset') still returns the ModelList or object array.

Fix getColumn([m, n]) It was returning child 1 (index of n) rather than child n.

Remove all traces of aria. SRs read better without Sad state of affairs, but NVDA and VoiceOver read the table better when no aria roles are assigned. Ticket #2531426

Sanitize the column _id for use as className For now, replace whitespace with dashes. There may be other cases requiring further cleanup.

Sanitize the column _id for use as className For now, replace whitespace with dashes. There may be other cases requiring further cleanup.

_initRecordType optimz, -arrow flip on col hover

Note the addition of _yuid, _id to columns API doc update.

+aria-readonly, role=presentation to templates Core defines the grid role on the contentBox (should this be on the boundingBox?) so all <table>s should be role presentation. No need to move ARIA roles around.

table.aria-describedby = caption

Don't be too clever with caption It was checking that _captionNode was attached to the table, but that was getting in the way of x scrolling. No need to be so pendantic. :)

Remove name-based fallback in getColumn Column map is keyed by key or name, so it's unnecessary. And an infinite loop because the if (i > len) should have been if (i >= len).

Add delegate method in its current implementation for backward compatibility. Decision made in a rush, so I may regret it.

Add headers to nested header `<th>`s Screen readers still hate my markup for this, but we'll get it worked out soon.

s/addAttribute/setAttribute/ oops

Add grid role to contentBox, presentation to table Odd setup for a static table, but the assumption is that no matter how the table is rendered, it should always be considered a single table to screen readers.

Review feedback #3 (from Satyam) * typo in getRow * move caption and summary attr => _uiSet* to _UI_ATTRS * rename {classes} placeholder {className} * header cells get column class as well * header cells get yui3-datatable-header class * data cells get yui3-datatable-cell class * Core now assigns columns a _id which is name, key, or yuid w/ uniqueness counter * column class names now yui3-datatable-col-foo (added -col-) * use toJSON() to get data for the formatters * replace CSS targeting tags with classes

Review feedback round #1 1. Fix typo in TFOOT_TEMPLATE 2. Don't store the array data populating the `data` ModelList in the attribute 3. Other minor typos

Remove backslash escaping in `code _snippets` Not needed in the latest yuidoc

-column widths (extension). +data bubbles to table Extracted the column width assignment code to the datatable-column-widths module. Made the ModelList bubble events to the DataTable instance.

Refactor the rendering mechanism to use events The view instances were inaccessible prior to render(). So I moved the renderUI logic into a renderTable event defaultFn. That defaultFn then sets up the table, caption, and columns, then fires renderHeader, renderFooter, and renderBody events. Each of these events call the view's render() method and finish up. Flexibility achieved!

_uiSetCols takes boolean to force col replacement Also, minor bug fix - remove subscription to widthChange. _afterWidthChange wasn't defined, resulting in the object syntax subscription to after() to cause widthChanges to be handled by _afterSummaryChange. Oops! Widget already wired up the widthChange sub, so no need to subscribe.

Add colgroup,col rendering and col width support Column widths are assigned via CSS on the <col> node associated with a given column, which is effectively a min-width. If content in the column cells will not fit in the configured width, it will expand the column. This is unfortunate native DOM behavior. I will create a separate module that enforces column width-based truncation, but that's not core functionality.

Move tbody,foot,head creation into Core It simplifies logic to have the Views working with the table sections rather than the table itself. Unrelated, renamed _uiUpdateWidth to _uiSetWidth to override the Widget method that just set dims on the boundingBox.

width attribute now sets table width

getColumn more versatile; get('columns.foo') fixed

Move `renderUI` to @protected section

Split out view config properties Class extensions can add to the config object passed to the constructor of the appropriate view when the render() process instantiates the view.

Allow `null` for views. API doc clarifications.

s/validator/setter/ for recordType The setter method was already in place, I just forgot to update the ATTRS.recordType config.

Finish API docs for current API There will be more. Oh, so much more.

Add API docs and other minor updates.

Add API docs and minor fixes as I see them Ongoing.

Add `getCell` and `getRow` pass thrus Also pass the `_cssPrefix` to the views for use by their `getClassName` implementations, and instantiate the view classes with `container` and `modelList`

Decouple column parsing, +nested cols The parsing of column array data was in Core, but was parsed into three different forms: 1) for column access from Core's get('columns.<key>') 2) for column prep for header rendering 3) for column prep for body rendering Now the columns are passed to the views to do with as they will, and Core only does what is needed for #1. Also, parsing nested headers efficiently is a pain in the ass. The 'children' config doesn't allow implementers to choose which cells get rowspan when the number of header rows is greater than the number of column descendant levels. As is done in 3.4.1, parent cells are always rowspan 1. I prefer to have data related columns have rowspan 1, but the logic to support that with the "children" config would be crazy complex. Implementers can move stuff around in `table.head.columns` before `render()` if they want different cell placement and spanning.

+getCell and getRow. Grouped, alphabetized methods Conflicts (resolved) src/datatable/js/pe.js

+ column.formatter & column.nodeFormatter Other bug fixes. Can render from array. Column ordering works. Movin right along.

Initial drop of DataTable rearchitecture Renders a table from supplied columns and data as a ModelList with Model instances. A lot more plumbing in place, but not yet exercised. Progressive enhancement extension in src/pe.js, but doesn't have build files yet.