Preventing the reordering of columns from replacing the `<span> tags in normal datatables styling or the  `<a> tags used on each button if using bootstrap styling.  This patch fixes the first part of #26.

appending the popup information box to the table instead of the html body so it always opens above the table.

Updating logic to check for children elements.

updating fork

Patch 1

Update dataTables.buttons.js

updating

I have been playing around with the export features today, and I came up with several minor improvements which I think enhance the functionality of the buttons export feature by allowing users to customized how both the data and the headers are exported.  

The first change I recommend(but am not submitting as part of this request) is to change the default orthogonal type from 'display' to 'export' and add a note to the documentation that the buttons function extends the orthogonal types used by the ColumnDefs render so that people can pre-process the data with special formatting, characters, etc when exporting the data.

Secondly, and this is where my new pull request fits in. It would be nice to also be able to customize the column headers when exporting the data.  Currently the export function just uses the existing column headers which are shown on the table when exporting.  This pull request adds a configuration option to the exportOptions called "columnHeaders" which defaults to the current setup and just returns the column headers, but which can be configured by passing a function which will recieve the existing column headers and which can then return a processed version of the headers.  This is useful for internationalization where you want to change the language used in the titles, as well as when you are exporting data which needs to be formatted in a particular way for another application to properly read it.

I updated the function, and applied it to the footers also.  

The function is now called for each column which is being exported and it is passed 3 values.  The 'colnum' is the column number where the data is from, the 'title' is the existing display title and the default response, and the 'column' is the column object.  Whatever string/value is returned from the function after the function runs is then used to populate the column header/footer.

If someone creates their own parse function for the column headers they can use the columnHeaders config setting to replace the default function with their custom function to custom parse the header data.

Patch 3