Anchors and flags

An anchor matches a position rather than a character, which makes it the cheapest thing in a pattern and the easiest to get wrong: nothing appears in the match to tell you the anchor meant something other than what you assumed. A flag has the same property one level up — it changes what ^, $ and . mean without changing a character of the pattern.

Lesson	The question
$ is not the end of the string	$ matches at the end of the subject or just before a final newline in Perl, Python, Java and PCRE2 — and in Ruby ^ and $ are line anchors always, with no flag involved. ^\d+$ is a question about a line, not about a value, and a validator built from it can be walked past.
Flags, and where you can turn one on	Eight engines spell case-insensitivity i and then part company: m is multiline in seven of them and dot-matches-newline in Ruby. Inline (?i), scoped (?i:…), free-spacing (?x), and which engines take which.

The rule the chapter ends on: for validation, anchor with \A and \z, or use the language's own whole-string call — re.fullmatch, String.matches, an anchored \A…\z in Ruby. ^…$ is a line test that looks like a string test, and the difference is exactly one newline wide.

A collision worth memorising before it costs you a day. Python's \Z is Perl's \z. Perl's own \Z is something else again, and JavaScript has neither. That is three engines, one escape, and no error message anywhere.