如何等待内部承诺完成

时间:2019-07-15 17:12:33

标签: javascript node.js promise

我正在使用fs.readdir获取目录列表,然后再次在回调中获取每个目录中的“子页面”列表。我希望第一个回调等待第二个回调完成,但是我不确定该怎么做。

// Array to hold list of pages
const pageList = []

// Get contents of target directory (not recursive)
fs.readdir(targetDir, { withFileTypes: true }, (err, items) => {
  // Stop and return if error
  if (!!err) return err

  // Go through found contents
  const theseItems = items.map(item => {
    const subpages = []

    // Directory name
    const dirName = item.name

    // Set up published target for this directory
    const thisTargetDir = targetDir + '/' + dirName + '/publish'

    // Now get pages in the directory's published directory
    // (assumes all files within subdirectories are .mdx pages to load)
    return (
      fs.readdir(thisTargetDir, { withFileTypes: true }, (err, pages) => {
        const theseSubpages = pages.map(page => {
          const mdxSuffix = /.mdx$/g
          const pageName = page.name.replace(mdxSuffix, '')

          return subpages.push({ name: pageName })
        })

        Promise.all(theseSubpages).then(() => {
          // Add to page list array
          pageList.push({ name: dirName, subpages: subpages })
        })
      })
    )
  })

  Promise.all(theseItems).then(() => {
    console.log('pageList at the end is: ')
    console.log(pageList)
  })
})

Promise.all(theseSubpages)可以按预期工作,但是Promise.all(theseItems)在前者有机会循环通过之前就解决了。我了解为什么会发生这种情况,并且我尝试过做类似Promise.resolve()等将每个item返回的事情,但是这些事情没有用。

想知道我是否在用这种方法做一些天生的错误……

更新

我尝试使用fsPromises方法,但一直遇到相同的错误模式。最后使用node-dir包来递归浏览目录。下面的代码,并不是我要执行的操作的确切答案,但这可以得到我想要的结果。

const dir = require('node-dir')

const targetDir = __dirname + '/../pages/stuff'
const pageList = []

dir.paths(targetDir, (err, paths) => {
  if (err) throw err

  const baseMatch = __dirname.replace('/lib', '') + '/pages/stuff'

  paths.dirs.map(dir => {
    // Only publish paths
    if (dir.substr(-7) === 'publish') {
      // Get the slug directly before publish path
      const contentSlug = dir.split('/').slice(-2)[0]

      // Add this to main pageList array as top level objects
      pageList.push({ name: contentSlug, subpages: [] })
    }
  })

  paths.files.map(file => {
    const filePathArray = file.split('/')

    // Only publish paths
    if (filePathArray.slice(-2)[0] === 'publish') {
      // Get parent content slug for matching purposes
      const parentContentSlug = filePathArray.slice(-3)[0]

      // Get file name (remove .mdx suffix)
      const mdxSuffix = /.mdx$/g
      const fileName = filePathArray.slice(-1)[0].replace(mdxSuffix, '')

      // Loop through main page list, find match, then add file as subpage
      pageList.find((obj, key) => {
        if (obj.name === parentContentSlug) {
          return pageList[key].subpages.push({ name: fileName })
        }
      })
    }
  })

  console.log('pageList at end:')
  console.log(pageList)
})

2 个答案:

答案 0 :(得分:0)

通过链接.then调用(Promise.then(doStuff))来工作。如果您许下诺言但又不做承诺,那么您将无法知道它何时完成。为了从内部函数链接承诺,您必须返回承诺。

通常,您不想混合使用回调和Promise。

如果我要这样做,我将只使用诺言。

const readdir = (target, options) =>
  // returns a promise that resolves or rejects when the call finishes
  new Promise((resolve, reject) =>
    fs.readdir(target, options, (err, result) => {
      if (err) reject(err);
      resolve(result);
    })
  );

const collectSubPages = pages =>
  // Wait for all the promises in the array to resolve
  Promise.all(
    // for each page, return a promise that resolves to the page/subpage object
    pages.map(({ name }) =>
      readdir(targetDir + "/" + name + "/publish", {
        withFileTypes: true
      })
        .then(subpages => subpages.map(({ name }) => ({ name })))
        .then(subpages => ({ name, subpages }))
    )
  );


readdir(targetDir, { withFileTypes: true })
  .then(pages => collectSubPages(pages))
  .then(console.log);

答案 1 :(得分:0)

@David Yeiser,您可以使用Array方法.filter().map()以及如下的各种优化来更简洁地编写您自己的“更新”代码:

const dir = require('node-dir');
const targetDir = __dirname + '/../pages/stuff';
dir.paths(targetDir, (err, paths) => {
    if (err) {
        throw err;
    }
    const baseMatch = __dirname.replace('/lib', '') + '/pages/stuff';
    const mdxSuffix = /.mdx$/g; // define once, use many times
    const fileList = paths.files
        .map(fileName => fileName.split('/'))
        .filter(filePathArray => filePathArray[filePathArray.length - 2] === 'publish'); // Only 'publish' paths
    const pageList = paths.dirs
        .filter(dir => dir.substr(-7) === 'publish') // Only 'publish' paths
        .map(dir => {
            const name = dir.split('/').slice(-2)[0];
            const subpages = fileList
                .filter(filePathArray => filePathArray[filePathArray.length - 3] === name) // select those files whose "parent content slug" matches 'name'
                .map(filePathArray => filePathArray[filePathArray.length - 1].replace(mdxSuffix, ''));
            return { name, subpages };
        });
    console.log('pageList at end:');
    console.log(pageList);
});

您将看到:

  • fileList是用paths.files.map().filter()模式构造的。
  • pageList是用paths.dirs.filter().map()模式构造的。
  • 对于pageList中的每个条目,subpages都是用fileList.filter().map()模式构造的。

除非我有错误,否则应该给出相同的结果。

未测试