Regex captures not working

[blystii]

I tried the example on Lune's official docs page, even with the latest version that I just installed and it doesn't work

local Regex = require("@lune/regex")
local re = Regex.new("hello")

if re:isMatch("hello, world!") then
	print("Matched!") -- This works
end

local caps = re:captures("hello, world! hello, again!")

print(#caps) -- This outputs 0

[PhantomShift]

This is related to #233, as another case of how slightly confusing the captures API is, although it's unfortunate that this confusion has leaked into the docs.

regex:captures only gets the first instance of a Regex::Captures in the subject, which is the set of captures in a singular match. Take for example this script:

local Regex = require("@lune/regex")
-- Note that double backslashes are needed for this to parse correctly
local ID_REGEX = Regex.new("(\\w+)\\s+(\\d+)")
local test_input = [[
admin 0
bob 42
]]

local caps = ID_REGEX:captures(test_input)
print(#caps) --> 2, since there are two capture groups in the regex
print(caps:get(1)) --> "admin"
print(caps:get(2)) --> "0"

As you can see, only the first match is returned by captures, and as captures_iter isn't exposed, currently one has to write something like this for the example in the docs to work as expected. I also note in hindsight that any offsets in the capture objects after the first one in this script will be off.

Script

local Regex = require("@lune/regex")
-- Note that it's in parenthesis to actually create a capture group,
-- which is another issue with the example in the doc
local re = Regex.new("(hello)")

-- In-lua implementation of captures_iter
local function captures_iter(regex, subject: string)
	local s = subject
	local n = 1
	return coroutine.wrap(function()
		repeat
			local captures = regex:captures(s)
			if captures == nil or #captures == 0 then
				break
			end
			coroutine.yield(n, captures)
			n += 1
			s = s:sub(captures:get(#captures).finish)
		until captures == nil or #captures == 0
		return nil
	end)
end

if re:isMatch("hello, world!") then
	print("Matched!")
end

for i, caps in captures_iter(re, "hello world! hello, again!") do
	print(i) --> 1, 2
	print(#caps) --> "1"
	print(caps:get(1)) --> "<RegexMatch(hello)>" (Does not have a tostring implementation)
end

I wrote a very basic implementation for exposing captures_iter in that issue but I wrote that to be as basic and non-intrusive to the rest of the code as possible so it probably doesn't meet expectations.

Regardless, I do think that repeated matches is something you would commonly do when using regex so making it easier and less confusing would be nice.