Stuck for no reason?

[crapthings]

i have a list that has 1.8k urls, but when

await crawler.queue(urls)

it seems stuck randomly without timeout?

const fs = require('fs')
const _ = require('lodash')
const writeJsonFile = require('write-json-file')
const HCCrawler = require('headless-chrome-crawler')
const RedisCache = require('headless-chrome-crawler/cache/redis')
const cache = new RedisCache({ host: '127.0.0.1', port: 6379 })

let urls = getUrls()
let count = urls.length

async function p1() {
  const crawler = await HCCrawler.launch({
    cache,
    persistCache: true,

    evaluatePage: (() => ({
      title: $('#litZQMC').text(),
      html: $('#divScroll').html()
    })),

    onSuccess: async resp => {
      const { result: { title, html } } = resp
      if (fs.existsSync(`files/${title}.txt`)) {
        console.log('skip', count--, title)
      } else {
        await writeJsonFile(`files/${title}.txt`, html)
        console.log('done', count--, title)
      }
    },

    onError: err => {
      console.log(err)
    }
  })

  await crawler.queue(urls)
  await crawler.onIdle()
  await crawler.close()
}

async function queue() {
  await p1()
}

queue()

• Version:
  1.8.0
• Platform / OS version:
  osx
• Node.js version:
  v8.11.3

[bloody-ux]

I have the same situation, but not randomly, it just stuck with chrome process killed after several minutes.

[davidebaldini]

Did anyone find a solution or workaround?

No exception is thrown and no error is printed. I still have a few chrome processes running when the script gets stuck.

[davidebaldini]

I discovered what causes blocks in my case.

The blocks happen when a tab is pointed toward (_page.goto()) a page containing flash. There, the browser shows a warning dialog which is not detected by _handleDialog() in crawler.js, and causes an infinite delay in _collectLinks().

Solution (works for me):
the first part of _collectLinks() needs to be changed to:

  /**
   * @param {!string} baseUrl
   * @return {!Promise<!Array<!string>>}
   * @private
   */
  async _collectLinks(baseUrl) {
    const links = [];
      await Promise.race([
        new Promise(function(resolve, reject) {
          setTimeout(resolve, 10000);
        }),
        this._page.exposeFunction('pushToLinks', link => {
          const _link = resolveUrl(link, baseUrl);
          if (_link) links.push(_link);
        })
      ]);
      console.log("PASSED");

Possibly this modification causes a memory leak, but it worksForMe.

[BubuAnabelas]

Maybe @yujiosaka could look more into this since it's clearly an easy to reproduce and easy to fix bug.

[popstas]

Try to add args: ['--no-sandbox'] to crawler options.

[kulikalov]

is anyone willing making a PR?