I am using selenoid for automated e2e-testing and really like it. To ensure browser support on all versions, I am testing with different versions of Chrome/FF and now would need to also test with the beta versions of these browsers.
I found out here that there is no built-in image for beta versions.
Is there another way to use beta versions anyway?
We provide all automation scripts used to build browser images. See this repository.
Regarding beta images - we have an issue and hope to start building more versions including ESR and beta during next months.
Related
At our company, we maintain a legacy Android application which uses Crosswalk to embed a web view. We would like to implement automated end-to-end tests for this app using Appium and ChromeDriver.
Unfortunately, the version of Chrome used by Crosswalk is modified in such a way that ChromeDriver cannot connect to it. This is a known issue, and the recommended workaround is to use a custom patched version of ChromeDriver instead.
This solution works for running tests locally. But BrowserStack doesn't seem to allow for supplying a custom build of ChromeDriver. I can choose from a set list of versions, but that won't help, as none of those versions have the Crosswalk patch.
Is there a way to make this work with BrowserStack?
I had used BrowserStack for my native app tests as well. Currently, they do not provide any custom capability to select the ChromeDriver version.
I have created a docker selenium image on ubuntu and have my selenium test cases executed with it. But I do see quite a few random failures.
Does anyone know the good version of:
1) Chrome,
2) Chromedriver,
3) Selenium,
4) TestNG,
5) Docker(?)
6) JDK
that totally it may give me the highest success rate ? Thanks.
There is no all-purpose correct version combinations here. I believe that you can evaluate this trying continuously different versions and see which one of them behaves the best for you.
There will always be some random failures caused by setup issues.
1) Chrome
This depends on the application you are testing and how it behaves on the given browser version. Some changes to the browsers affect in a terrible way some web apps. So we can't propose you specific version without any knowledge about your app. It really depends on many factors.
2) Chrome Driver
Each version of Chrome Driver is recommended for specified chrome versions so it depends on the browser version.
3) Selenium has some bug fixes in it's newer versions so in the perfect case you should use the newest version. But there is no guarantee that it will fit you the best.
4) TestNG - it depends on the way you use it so nothing more can be said here.
5) Docker - not sure if the Docker version can affect in some way the test executions. But it might depend on the selenium version.
6) JDK depends on your tests implementation. Are you using new java features and etc. Actually it should have minimal impact over the test executions
I am struggling with a PhantomJS/GhostDriver issue in my CI builds. (Geb/Selenium tests hang loading new page)
Unfortunately there is no PhantomJS 2.0 binary for Linux yet, and I have not successfully been able to build it from source.
In the meantime, is there a way I can run the latest GhostDriver from within an older version of PhantomJS?
I've just read about the release of Safari4 (beta) and its much improved Javascript engine.
I was planning to start playing with it, but I am wondering if anybody knows if both Safary 3 and Safari 4 can be run side by side.
You could run Safari 3 and use the nightly builds of Webkit to test the new engine. You'd be still missing the new features like Cover Flow and Top Sites though.
Try this
Safari normally use the Web Kit
framework found inside Mac OS X to
render web pages and execute
javascript. This means that if you
preserve an old version of Safari to
run it on a newer version of Mac OS,
it will use the newer Web Kit found in
the system and you will get the same
results as with the newer version.
However, the link has some stand alone versions that you can install.
There's also this which shows how to create your own stand alone version.
Beatnik Pad has a quick tutorial that shows how running Safari 4 Beta and Safari 3 side by side can be accomplished.
Is there a way to install all versions of all browsers on one machine?
I use virtual machines to do different browser version configurations. I'm assuming you want to test different version of IE, Firefox etc.
If you automate your testing at any point the virtual machines are also handy for having a way to store many different configurations for testing on one machine.
Sure, and the best way to do this is to use virtualisation technology such as VirtualBox.
we have used this to install mutiple versions of IE for testing it goes as far back as 3.0
http://tredosoft.com/Multiple_IE
Try Sandboxie. Much less work to deal with than a VM. Probably will run faster.
For several browsers, like Firefox or Opera, it is just a matter of copying the original program directory elsewhere and upgrading. Most of the time, the old version should still work.
For IE, there are several distributions, IETester seems to be a complete, easy solution.
For FireFox, you can install the portable versions available here
With IE, I think the only way is via virtual machines. I can recommend VirtualBox, works smoothly for me.
Why would you want to do that?
Typically, a product wont allow parallel installs (multiple versions of same product, though certainly you can have multiple versions of different products).
If you want to do some analysis or testing , doing that in separate Virtual Machines (check out VMWare) may be an option.
I am not sure why you really want this but if you want to install different versions of IE use the following tool:
http://tredosoft.com/Multiple_IE
Yes. I have IE8 Beta2, FF3, Chrome Beta, and Safari 3.1.2 with the applicable developer plugins.