I've recently found and fallen in love with Cypress for e2e testing, much more so than Selenium, but there's a catch: our web app will actually be within jxbrowser, inside a java app. Whatever e2e framework we choose has to be able to run all the tests within jxbrowser inside of this embedded application (hence why we originally looked at selenium, because of the remote driver)
Is it possible to get this working? We absolutely need the embedded jxbrowser tests (Cots integration, can't get around it), and I'd hate to have to fall back to selenium.
Related
I have a big web project with a separate backend and a front-end (webpack). I'm going to use Cypress to create end-to-end tests.
What is not clear is where I should add the Cypress tests and Cypress itself. The documentation says to add it right to the testee project and it shows how to run the tests on the production website (which URL is different from the local, dev project). This means that I'm not able to run the tests on the development project because Cypress testing IDE and the testee project can't be run simultaneously because they share the same terminal.
If so, the best solution is probably to organize one more project, only for testing purposes, and having only Cypress installed and tests themselves? Is it a good practice and if so, which project should it be?
We have the same setup at work. We include the Cypress folder in the front-end repo. I'd agree with keeping it right next to the project because you have access to that code easily i.e. accessing utility functions, selectors, etc. As far as the terminal issue, you should be able to run your project locally in one terminal tab and the cypress test runner in another.
I am currently working on node-webkit based desktop app. I was searching for a way to do automation testing for this . But most of the automation tools required a server to start up but in my case i have on but it runs in the chromium that comes by default. Is there a way to automate UI acceptance test for nwjs apps.
Take a look at the NW.js Wiki under How to run node webkit's test cases
You will need to setup the chrome driver in order to use selenium for example
Once you have those, check out Yaniv kessler's blog post on how to Automate NW.js UI testing using mocha.
I have some Selenium IDE scripts for automating tasks (deploying to an internal server via it's web interface), and I want to have bookmarks to launch them.
Don't want go to Tools -> Selenium IDE -> search test -> launch test, just want to click a bookmark and let the script run.
How can I achieve this? Am I missing a point?
Don't know if it makes a difference, but I'm currently using Firefox (if I have to change browser for this, it wouldn't be a problem).
Thanks
Not sure whether you be able to run tests pressing on a bookmark as browser plugins are usually run with a key combo. And bookmark is URL. Well the main issue in your question how to associate bookmark
with browser addon launching.
Well, I've investigated a lil bit workaround. Speaking about getting Selenium HTML Test Suites Running Automatically (i mean an alternative to bookmarks) some solutions are possible:
solution 1:
Patch the Selenium server code to support HTML suites.
Build a script to execute your HTML test suites.
Build a script to read the suites’ output reports and consolidate
the information for inclusion into your CC.NET report.
Setup details you can get here
Solution 2:
Selenium IDE plugin for Firefox to create tests
Java to run Selenium Server
Selenium Server to run test suites
Nant to generate reports for CruiseControl.NET
CruiseControl.Net as a continuous integration engine
Setup details here
Hope this be useful for you.
Bookmarks are nothing else than URLs, which makes it impossible to "bookmark" a functionality of a plugin, such as the Selenium IDE.
However, there might be other plugins out there that do what you want. I haven't had a closer look at it, but one candidate could be the Favorites (Selenium IDE) Firefox Add-on.
Selenium: Is there any JS (JavaScript) code coverage tool which we can integrate with Selenium Server/RC
Hi,
I want to JS code coverage done by my Selenium Test Suite
Is there any code coverage tool which can be integrated with my existing selenium rc framework??
In theory, you can use any coverage tool that instruments JavaScript code to collect coverage data.
For example, you can use JSCoverage either in proxy mode (for real-time instrumentation) or for static instrumentation to run your Selenium tests off the instrumented code.
One technical hurdle could be to devise a way to accumulate coverage data across multiple pages. The coverage tool of your choice could already have support for accumulation of data from multiple page runs. If not, you might have to handle that bit yourself, but this is probably not something trivial.
I have found the solution for my need, there are multiple tools available but I want a tool which can be integrated with my existing automation easily
Tool is JSCover
http://tntim96.github.io/JSCover/
Just run the Server
Run Server
java -cp JSCover-all.jar jscover.server.SimpleWebServer . 8080
Run Proxy Server
java -jar target/dist/JSCover-all.jar -ws --proxy --port=3128 --report-dir=target
Add Proxy in your Webdriver for port specified, 3128 in my case
After execution of test Generate Reports using
((JavascriptExecutor) driver).executeScript("jscoverage_report();");
You can add this line where you exit WebDriver or Test Case
Not aware of a tool for Selenium, but JsTestDriver has a design very similar to Selenium RC (can launch tests from the command line and they are run on a server that drives browsers headlessly) and provides code coverage information.
IntelliJ integrates with JsTestDriver and provides a visual display of coverage information.
Take a look at https://github.com/yui/yuitest/wiki/Code-Coverage
I've integrated this tool in my Selenium tests a time ago. You need a bit of work to gather coverage info before page changes (in any case js trigger a page reload, link etc...)
Once you set up everything, it will fully coverage any js executed while Selenium load and test your website pages.
PS : Even if it was specially adapted for YUI test, you can use it with selenium.
I don't know what you are trying to achieve, but:
Selenium is testing the final output, as seen on the page itself.
So it really does not matter if its PHP, HTML, JSP, ASP or .NET - the Selenium is designed to mimick the end user and click the final application - the final HTML code generated by whatever is under the hood.
Selenium is also not that good for code coverage tests - one piece code can be on many pages - so better approach with selenium is to do the "user" coverage - try to cover all the possible actions which living human could possibly do with your page
There is no particular tool that can integrate with Selenium to do JS coverage. However there are lots of tools which test JS on every page which can tell if the JS that executing on your web page had any errors. This may not ideal solution but on each page you will have the measure of uptil which point JS executed properly on your webpage under test. There are two solutions for that:
1.) JSErrorCollector API: It will integrate directly with Selenium and let you know if there were any error on the page. Can be found at: http://mguillem.wordpress.com/2011/10/11/webdriver-capture-js-errors-while-running-tests/
2.) Full fledged JS coverage tools: There is an excellent list of tools here which will essentially help you in covering JS on on your web pages. Can be found at: JavaScript unit test tools for TDD
I'm writing functional tests for a chrome app written entirely in javascript. I decided to use Cucumber and Selenium Webdriver and run the tests on chrome. Wrote Selenium code inside Cucumber step definitions and it works fine (I've only just begun though).
I see that I can do headless testing with capybara (or webrat) and I'm not interested in that. Is there anything else that I'd gain if I use capybara?
I recently had the same question. Capybara gives you a pretty nice DSL to interact with the DOM for one thing. Another big advantage is you can easily switch drivers. On my team, we constantly switch from the Chrome driver (desktop browser) to the iWedbDriver (mobile safari in iphone simulator).
Headless testing is generally a very good thing, since the tests run faster if they don't have to draw the browser window. If the tests run faster, then you'll run them more often. If you're not interested in that, you're making a big mistake.