问题
I'm testing a site with lots of proxies, and the problem is some of those proxies are awfully slow. Therefore my code is stuck at loading pages every now and then.
from selenium import webdriver
browser = webdriver.Firefox()
browser.get("http://example.com/example-page.php")
element = browser.find_element_by_id("someElement")
I've tried lots of stuff like explicit waits or implicit waits and been searching around for quite a while but still not yet found a solution or workaround. Nothing seems to really affect page loading line browser.get("http://example.com/example-page.php"), and that's why it's always stuck there.
Anybody got a solution for this?
Update 1:
JimEvans' answer solved my previous problem, and here you can find python patch for this new feature.
New problem:
browser = webdriver.Firefox()
browser.set_page_load_timeout(30)
browser.get("http://example.com/example-page.php")
element = browser.find_element_by_id("elementA")
element.click() ## assume it's a link to a new page http://example.com/another-example.php
another_element = browser.find_element_by_id("another_element")
As you can see browser.set_page_load_timeout(30) only affects browser.get("http://example.com/example-page.php") which means if this page loads for over 30 seconds it will throw out a timeout exception, but the problem is that it has no power over page loading such as element.click(), although it does not block till the new page entirely loads up, another_element = browser.find_element_by_id("another_element") is the new pain in the ass, because either explicit waits or implicit waits would wait for the whole page to load up before it starts to look for that element. In some extreme cases this would take even HOURS. What can I do about it?
回答1:
You could try using the page load timeout introduced in the library. The implementation of it is not universal, but it's exposed for certain by the .NET and Java bindings, and has been implemented in and the Firefox driver now, and in the IE driver in the forthcoming 2.22. In Java, to set the page load timeout to 15 seconds, the code to set it would look like this:
driver.manage().timeouts().pageLoadTimeout(15, TimeUnit.SECONDS);
If it's not exposed in the Python language bindings, I'm sure the maintainer would eagerly accept a patch that implemented it.
回答2:
You can still speedup your script execution by waiting for presence (not waiting for visibility) of expected element for say 5-8 sec and then sending window.stop() JS Script (to stop loading further elements ) without waiting for entire page to load or catching the timeout exception for page load after 5-8 seconds then calling window.stop()
Because if the page not adopted lazy loading technique (loading only visible element and loading rest of element only after scroll) it loads each element before returning window.ready state so it will be slower if any of the element takes longer time to render.
来源：https://stackoverflow.com/questions/10750198/set-up-a-real-timeout-for-loading-page-in-selenium-webdriver