Why do ^ and $ not work as expected?

This puzzled me the last 15 minutes:

if ('ab' =~ /^a|b$/) { print 't' } else { print 'f' }
print "\n";

I have expected that 'a' or 'b' following the beginning and followed by the end, should match only one character. So the test should fail for two characters 'ab'. But it succeeds. Why?

If you group the alternation, then you will get the expected behavior:

/^(a|b)$/

Your regex will find a a at the start of the string (with ^a branch) or b at the end (with the b$ branch).

When you use ^(a|b)$, the anchors are applied to the whole group and thus it will match a string that is equal to a or b.

Also, if you do not really need to capture the value, you may either use a non-capturing group, /^(?:a|b)$/, or a n modifier, /^(a|b)$/n.

Your regexp matched ^a or b$, because the alternative operator | has lower precedence than a sequence of concatenated regexps.

In this particular case (an alternative of single characters) you can simplify it to a class:

/^[ab]$/