【发布时间】:2013-02-05 18:14:00
【问题描述】:
我需要用 attoparsec 解析固定长度的字段,但我现在正在努力使用编译器。 我还是个新手,下面的代码是我最接近的解决方案:
> {-# LANGUAGE OverloadedStrings #-}
> import Control.Applicative
> import Data.Text as T
> import Data.Attoparsec.Combinator
> import Data.Attoparsec.Text hiding (take)
> import Data.Char
> import Prelude hiding (take)
>
> type City = Text
> type Ready = Bool
> data CityReady = CR City Ready deriving Show
>
> input = T.unlines ["#London 1",
> "#Seoul 0",
> "#Tokyo 0",
> "#New York 1"]
>
> parseCityReady = many $ CR <$> cityParser <*> readyParser <* endOfLine
>
> cityParser = char '#' *>
> takeTill isSpace <*
> skipWhile isHorizontalSpace
>
>
> readyParser = char '1' *> pure True <|> char '0' *> pure False
>
> main =
> case parseOnly parseCityReady input of
> Left err -> print err
> Right xs -> mapM_ print xs
>
这一切都很好,但它只是返回没有空格的城市。
CR "London" True
CR "Seoul" False
CR "Tokyo" False
我尝试使用 applicative 为 City 文本字符串取 20 个字符
> cityParser = char '#' *>
> take 20
甚至do syntax
> cityParser = do char '#'
> city <- take 20
> return city
但两种方法都无法编译并出现此错误:
Couldn't match expected type `attoparsec-0.10.4.0:Data.Attoparsec.Internal.Types.Parser
Text b0'
with actual type `Text -> Text'
In the return type of a call of `take'
Probable cause: `take' is applied to too few arguments
In the second argument of `(*>)', namely `take 20'
In the expression: char '#' *> take 20
当take 的类型为Int -> Text -> Text 时,是什么导致ghc 请求Text -> Text?
如何在 applicative 和 do-syntax 中解决它?
【问题讨论】:
-
您隐藏了由
attoparsec导出的take的版本。相反,您使用的是由Text模块导出的take函数,这不是您想要的。 -
@sabauma 确实,我应该将
hiding放在 Data.Text 上。谢谢! -
@sabauma:如果这是正确答案,您应该发布它
标签: haskell monads applicative attoparsec