Combining marks and accented letters behave as expected: Your naïveté amuses me!
Unbreakable: $9 $(12.35) $[9] $-9 (12)¢ 12.54¢ -1/2 99% [9]% 1,234 (s)he file(s)
Breakable: "()()" "$[" "]%" "/9"
`++` and `--` are operators in C++.
Hyphenated words are breakable: server-side
Em dashes can be broken on either side, but not in the middle: A——B
NOT each space is breakable, like the last one in the following sentence (in English): He did say, “He said, ‘Good morning.’ ”
Regretfully it does not work for the alternative quoting style due to the ambiguity of “’” (quote and apostrophe): one should not insert normal spaces between the last two quotation marks in ‘He said, “Good morning.”’ — Of course, non-breaking spaces are OK: ‘He said, “Good morning.” ’
Still, the space rule makes my favourite ellipsis style work well, like ‘once upon a time . . .’.
Here's an apostrophe; and there’s a right single quotation mark.
Spanish inverted punctuation marks are always opening (e.g.)¡Foo bar!
There should be no break after Hebrew + Hyphen like א-ת, or between solidus and Hebrew like ‘/ת’.
Lines like this should not have their punctuation broken!… 
Emojis joined by ZWJs should not be broken: 👩‍❤️‍👩.
Breaks can occur between Chinese characters, except before or after certain punctuation marks: 「你看過《三國演義》嗎？」他問我。
