Could still be some corner cases since I kind of wrote this literally.
(.{1,%d})( +|$\\n?)|(.{1,%1$d})
consists of three capturing groups, the match is either the first two OR the third one
first: (.{1,N}) The %d is replaced with the width so this captures from 1 to N characters which are not newlines
second : ( +|$\n?) [extra \ is from string display] capture 1 or more spaces (there is a space before the +) OR end of the line with optional newline. Had to make newline optional to handle 1 line strings such as this one. (that is a hard one to catch at first)
third: (.{1,N}) grab the first N non-newline characters
First choice is to match 1 to N characters which are followed by spaces, the end of the string or newline. This is greedy so it gets the longest possible match.
If the first choice fails, you simply grab the first N characters
Replace: Here is another one of my favorite tricks. There are two capture groups we care about either #1 or #3. One will match, the other is empty and we want to replace with the one that matched. If you replace one followed by the other ($1$3) then either way you get the string that matched. Simply add a newline to do the word wrapping. Hard part is to remember to start counting at 1.
