1

我正在玩 Haskell 中的Channels。我想IO并行执行一些操作，直到其中一个失败，然后将所有结果收集到一个列表中。

此代码与Exception <<loop>>. 我怎样才能让它工作getChanContents？我见过的所有示例都假设他们知道频道上有多少消息。

有没有更干净的方法从工人那里收集一堆结果？

module UrlPatterns where

import Control.Concurrent
import Types
import Text.HTML.Scalpel
import Data.Monoid ((<>))
import Control.Concurrent.Chan
import Control.Applicative
import Data.Maybe (isJust, catMaybes)
import Data.List (takeWhile)


-- find all valid links under a domain that follow the pattern:
-- http://example.com/pages/(1..N)
-- as soon as one is missing, return a list of all the ones you found
findIncrementing :: URL -> IO [Link]
findIncrementing base = do

    let num = 1

    -- find channel
    cfind <- newChan
    writeChan cfind (base, num)

    -- results channel
    cdone <- newChan

    forkIO $ worker cfind cdone

    -- collect the results
    results <- getChanContents cdone
    let results = takeWhile isJust results :: [Maybe Link]
    print results

    return []


worker :: Chan (URL, Int) -> Chan (Maybe Link) -> IO ()
worker next done = loop
  where 
    loop = do
      (base, num) <- readChan next
      let url = pageUrl base num
      putStrLn $ "FETCHING: " <> url

      mt <- findPageTitle url

      case mt of
        Nothing -> do
          writeChan done Nothing
          putStrLn ("Missed " <> show num)
        Just t  -> do
          writeChan done $ Just $ Link url t
          writeChan next (base, num+1)

      loop

scrapeTitle :: Scraper String String
scrapeTitle = text "title"

findPageTitle :: URL -> IO (Maybe String)
findPageTitle url = scrapeURL url scrapeTitle

pageUrl :: URL -> Int -> URL
pageUrl base num = base <> show num
4

1 回答 1

2

感谢@bartavelle。我有一个与频道代码无关的错误。这是相关的修复：

-- collect the results
results <- getChanContents cdone
let links = catMaybes $ takeWhile isJust results

return links
于 2015-04-08T17:07:53.423 回答