As described in #877 by MartinDawson:

In our spreadsheet we use a contentEditable div. We parse the text using chevrotain's lexer to find the cells to highlight. Like so:

When we access the content using .textContent it seems to replace all spaces with non-breaking spaces.

This should be fine because hyperformula should replace ANY space. However it's missing non-breaking spaces from the regex:

pattern: /[ \t\n\r]+/

Hyperformula's lexer was returning #ERROR! due to this.

So it should be /s+ instead.