Blank column headers throw cryptic error

a Zendesk user reported an unexpected message,

Error: TypeError: t is null

In this case, that error originated in CSV.js where we are now trimming whitespace from column headers, because this user had an extra column with a null value in the header position.

We should protect against this error, perhaps by replacing any null column headers with variants of "untitled". We must replace with something, because in this part of the code, the header row is being prepared as the keys used to reference row values in a JS object representing each parsed row -- simply removing the null without removing it from every data row would lead to other problems.

Start in src/js/core/CSV.js at the column-header trimming logic around line 86, and trace how header values become keys for parsed row objects. Handle null or blank headers without changing the alignment between headers and data rows; done means the parser produces usable keys and no longer throws the reported TypeError.