I close the issue because I posted the solution and did not wait for feedback for several weeks.

I close the issue because I posted the solution and did not wait for feedback for several weeks.

I close the issue because I posted the solution and did not wait for feedback for several weeks.

What is the status of this issue?

What is the status of this issue?

@voile

Invalid regex handling is not specified, e.g ?invalid, *invalid and +invalid. They don't even satisfy the grammar provided by the kata description, and still expects some matches, so what is expected?

OK, now expressions like /?invalid/, /*invalid/ and /+invalid/ invalid and must raise error.
I've changed description.

Changed to:

"[0-9]{3} USD"

Fixed. Thanks!

Fixed. Thanks!

Why?

(abc)+ expects ["abc", "abcabc", "abcabcabc"]

a+ expects ["a", "aa", "aaa"]

Am I right?
Why is it wrong with numbers?

Changed to:

• generator must have #each method which return Enumerator of strings (order doesn't matter)

Fixed. Thanks!

/(?<word>[a-z]+)|(?<number>\d+)/

Alternation are only supported within a group!

Notably, groups aren't dropped from the result just because they don't participate in the match, but the kata expects them to be dropped, which screws over whoever using these groups, as now the index of the group can change.

I've added tests for this cases.

        describe "named groups" do
          describe "names and alternation" do
            include_examples "#scan", "signed", {
              regexps: [
                "(?<word>[a-z]+)",
                "(?<number>\\d+)",
                "((?<word>[a-z]+)|(?<number>\\d+))"
              ],
              cases: {
                "abc 123" => {
                  "(?<word>[a-z]+)" => [["abc", 0..2, [], {"word"=>0..2}]],
                  "(?<number>\\d+)" => [["123", 4..6, [], {"number"=>4..6}]],
                  "((?<word>[a-z]+)|(?<number>\\d+))" => [
                    ["abc", 0..2, [0..2], {"word"=>0..2}],
                    ["123", 4..6, [4..6], {"number"=>4..6}]
                  ],
                }
              }
            }
          end

          describe "nested groups" do
            include_examples "#scan", "signed", {
              regexps: [
                "(\\d+)([a-z]+(\\d+))?",
                "(?<first>\\d+)(?<second>[a-z]+(?<third>\\d+))?",
              ],
              cases: {
                "1 1b 1bc2" => {
                  "(?<first>\\d+)(?<second>[a-z]+(?<third>\\d+))?" => [
                    ["1", 0..0, [], {"first"=>0..0}],
                    ["1", 2..2, [], {"first"=>2..2}],
                    ["1bc2", 5..8, [], {"first"=>5..5, "second"=>6..8, "third"=>8..8}]
                  ],
                  "(\\d+)([a-z]+(\\d+))?" => [
                    ["1", 0..0, [0..0], {}],
                    ["1", 2..2, [2..2], {}],
                    ["1bc2", 5..8, [5..5, 6..8, 8..8], {}]
                  ],
                }
              }
            }
          end
        end

For expression (\d+)([a-z]+(\d+))? expected groups with overlapsed ranges - [5..5, 6..8, 8..8].
For named groups with alternations expected hash - {"first"=>5..5, "second"=>6..8, "third"=>8..8}

The order of the results returned by scan isn't specified.

I've added this line to description

the order of the matches doesn't matter as long as they don't violate the principles described below

Yes, agree.
I've fixed solution and test cases.