{-|
Description:    

Copyright:      (c) 2020 Samuel May
License:        MPL-2.0
Maintainer:     ag.eitilt@gmail.com

Stability:      experimental
Portability:    portable
-}
module Test.Mangrove.Unit.Tokenize.Tag
    ( tests
    ) where


import qualified Data.ByteString.Char8 as BS.C
import qualified Data.Text as T
import qualified Data.Tuple.HT as U.HT

import qualified Test.HUnit as U

import Test.HUnit ( (~:), (~?=) )

import Web.Mangrove.Parse.Tokenize


tests :: U.Test
tests = "UTF-8 encoding" ~: U.TestList
        [ "appropriate end tags allow re-entry" ~: U.TestList $ map stepBreakpoint
            [ RawTextState
            , RCDataState
            , ScriptDataState
            , ScriptDataEscapedState
            ]
        ]


stepBreakpoint :: CurrentTokenizerState -> U.Test
stepBreakpoint state = show state ~:
    step' ~?= ([([], EndTag tag)], tokenState', BS.C.pack "</xmp>")
  where step = tokenizeStep tokenState $ BS.C.pack "</script></xmp>"
        step' = U.HT.mapSnd3 (tokenizerEncoding (Right $ Just Utf8)) $ step
        script = T.pack "script"
        tag = emptyTagParams
            { tagName = script
            }
        tokenState = tokenizerStartTag Nothing script $
            tokenizerMode state defaultTokenizerState
        tokenState' = tokenizerMode DataState tokenState