0.3.1 • Published 7 years ago

chromedriver-html5-dragdrop v0.3.1

Weekly downloads
523
License
MIT
Repository
github
Last release
7 years ago

Chromedriver HTML5 drag and drop fix

<text>

Problem statement

If you're using the HTML5 drag and drop implementation in you web app and want to test its behaviour with Selenium and Chromedriver, you're out of luck due to this bug:

Chromedriver can handle clickAndHold, move and release, but has a bug in this. There is a event.button attribute in every Javascript MouseEvent object. This is set to -1 if nothing is clicked and something between 0 (main mouse button) and 4 if a button is clicked. Now chromedriver sets this on clickAndHold but does not take it over into the move so when moving the event.button is -1 again.

Solution

Using Selenium's JS API, possibly wrapped by something like Nemo, simply include dragdrop-chromedriver.js in your test runner or -suite, like this:

    require('chromedriver-html5-dragdrop')(yourInstantiatedWebdriver/* , waitMillisBetweenSteps */);

When using Nemo, yourInstantiatedWebdriver is usually nemo.driver.

You can set waitMillisBetweenSteps for debugging or watching what's happening.

Now, use dragAndDrop just like in Selenium's example. Refer here for the dragAndDrop method's API.

Caveats

Because for security reasons synthetic drag events do not have a usable dataTransfer property, all dataTransfer accesses in the tested code must be made conditional, for example like this:

    if (event.dataTransfer) { event.dataTransfer.dropEffect = 'move'; }

Because of this, but also due to limits of the original dragAndDrop API, you won't be able to test dragging files into the browser window.

How it works

Basically it monkey patches webdriver.ActionSequence, overriding its dragAndDrop function with one that remains fully API compatible to the original one, but internally triggers the dragstart, drag, dragover, drop and dragend events in the browser context. There are some checks to prevent this approach from performing drag and drop operations a real user wouldn't be able to perform.

0.3.1

7 years ago

0.3.0

7 years ago

0.2.0

8 years ago

0.1.5

8 years ago

0.1.4

8 years ago

0.1.3

8 years ago

0.1.2

8 years ago

0.1.1

9 years ago

0.1.0

9 years ago

0.0.2

9 years ago

0.0.1

9 years ago