A fork of node-huxle, which is - A port of the codeless front-end testing tool, Huxley, used by Instagram.
- Records your actions as you browse.
- Takes screenshots.
- Compares new screenshots against the old ones and checks for differences.
- Allows executable JavaScript (non-recorded) code in your Huxleyfile.json. Example:
{ "action": "eval", "code": "do.something('on').then($.fn.blah)" }
- ?
npm install -g muxley
Selenium Server is used to automate the recorded browser actions. If you already have it, skip this. Don't have it and don't want the hassle of managing it? Download the node wrapper instead.
hux -h for a list of the available commands.
The API's so short it didn't warrant its own section. This example covers every feature of Huxley. If you have any Windows or browser issue during the process, see the FAQ below.
In examples/ you'll find two completed Huxley tests; we'll reproduce them below.
Let's start examples/ from scratch by removing everything but the webroot/ folder. We'll be testing component.html. Take a look.
cd into webroot/ and start a local server. Try python -m SimpleHTTPServer (if you're on Python 3.x: python -m http.server) or use this package (at port 8000).
Back in examples/, create a Huxleyfile.json, like this:
[
{
"name": "toggle",
"screenSize": [1000, 600],
"url": "http://localhost:8000/component.html"
},
{
"xname": "type",
"url": "http://localhost:8000/component.html"
}
]Each task is an object. Only name and url are mandatory and screenSize is the only other option. Note that the second task is marked as xname. This means that the task will be skipped. We'll focus on the first one for now.
Start Selenium (see "Installation" above), it doesn't matter where. Now run hux -r to start recording. The default Selenium browser is Firefox. Assuming Selenium started correctly, do the following:
- Go to the terminal, press
enterto record a first browser screenshot. - Go to the browser, click on the text field and type 'Hello World'.
- Back to the terminal again, press
enterto take a second screenshot. - Back to the browser, click on the checkbox.
- Back to the terminal one last time,
enterto take a third screenshot. - Now press
qenterto quit.
Huxley will then replay your actions and save the screenshots into a folder (notice that your actions are chained one after another without delay for a faster test experience).
Onward!
When you do need delays between actions for animations or AJAX, you can activate a special switch.
Open Huxleyfile.json again. Change the first task's name key into xname and the second one's xname into name (i.e. we'll skip the first task and run the second one). Run hux -r again:
- Go to the terminal,
lenter. This will capture a screenshot and mark a live playback point. From this point to the nextenter, Huxley will respect the delay between your actions. - Browser: click on
Launch modal. - Terminal:
lenteragain, as we'll be dismissing the modal and want to take the screenshot only after the transition's done. - Browser: click anywhere to dismiss the modal.
- Terminal:
enterto take a regular screenshot (thus marking the end of earlier's live playback), thenqenterto quit.
That's it! Check your replay and Don't forget to remove the x in xname of Huxleyfile.json.
When you modify your code in the future, hux to compare the new screenshots against the old ones and hux -u to update them. If you want to batch run Huxleyfiles, see hux --help.
Node-huxley is a port, so every recommendation here still applies.
Currently, only Firefox and Chrome are supported. For Chrome, you need chromedriver, which doesn't come bundled with Selenium yet. If you're using Brew, just do brew install chromedriver.
Start that then do:
hux -b chrome -flagForWhateverElseYouNeed
Make sure that:
- Java is installed and in your environment path.
- If the enter key doesn't register while recording, try typing anything (beside the reserved
qorl) before pressing enter. - If you're using the
--fileflag, use only forward slashes (/) in your pattern.