The bad news: from Firefox 55 onwards, Selenium IDE will no longer work.
The reasons for this are complex, but boil down to two main causes:
- Browsers are complicated pieces of software that are constantly evolving. Mozilla has been working hard to make Firefox faster and more stable, while still retaining the flexibility and ease of extension that we’ve come to know and love. As part of that process, Firefox is switching extensions from the original “XPI” format, to a newer, more widely adopted “Web Extension” mechanism.
- The Selenium project lacks someone with the time and energy to move the IDE forwards to take advantage of the new technologies.
Selenium is one of the most widely used pieces of testing software there is. Despite this, the team of people regularly contributing is small: since the start of the year, there are only 11 people who have made more than 10 commits, with two people accounting for more than half of those. Since 2016, only one person has been maintaining the IDE.
Selenium is an Open Source project. None of the core contributors — not the IDE maintainer, not the language binding owners — are paid to work on work on it. They do it because they love working on the code, and they typically do it in their “copious free time”. The IDE maintainer has had almost none of that to spare. We should all be thanking that committer for his time and effort. Thank you, Samit!
So what can we do to move forward? The first thing is that there are now a wealth of tools that are stepping up to fill the gap. You should go and have a look at them. The second thing is that there is an effort to rebuild IDE using modern APIs, to be usable across more than just Firefox. The fine people at Applitools are helping with this effort.
The third thing? That’s you. You could help us.
If you believe that a friendly UI for quickly recording and playing back tests is a useful Open Source tool, then please come and join us! The main technical discussions are happening on the #selenium IRC channel. If you’d prefer Slack, you can join us on that too. Or there’s the ever useful selenium-developers mailing list. Come onboard. We’d love your help, and IDE is a wonderful thing to contribute to!
Official Selenium Blog 
It seems like Selenium is slowly dying, at least for Firefox. Geckodriver is still in an unusable state, and progress is not really happening. Very many companies are stuck with Firefox 47 and have no clue how to proceed with Firefox test automation. The end of Selenium IDE is another step in this direction. Since a long time Chromedriver is much more reliable and stable compared to all the Firefox stuff. And Chrome develops its own test automation approaches which are independent of Selenium. Maybe it is time for a new approach to Browser automation.
Comment by Lars Frantzen — August 11, 2017 @ 7:44 am GMT+0000 |
I don’t think your information about geckodriver is accurate. With the latest releases, it’s at least as good as the old Firefox driver. The new Selenium 3.5 release contains a lot of work to make the standalone server work seamlessly with the w3c version of the protocol, which was one area of roughness.
The W3C “WebDriver” specification is nearing the end of its “Candidate Recommendation” period, and should become a standard soon. All the major browser vendors have their own implementations already up and running. The Selenium team have been involved with that effort too. When the spec is complete, a useful set of automation “atoms” will be in place for the next generation of browser automation to build upon.
Comment by shs96c — August 11, 2017 @ 9:00 am GMT+0000 |
Hi, I cannot confirm that the latest geckodriver + Selenium + Firefox combination is as good as the old Firefox driver. Whenever there is a new version of any of these I run our test suite against it, and also with the most recent versions the testing crashes or just halts. And it does not break or halt with Firefox 47 + Selenium 2.53.1. And this goes now for months, which is a huge hassle since you need security exceptions to keep the old Firefox running, etc.
You may now argue that we could file defects for all these reasons when it fails, but geckodriver started so very alpha that the motivation was low from the beginning. Though, the biggest issues seem to be solved in geckodriver now, but still it is not production-strength. At least not when you come already with an existing test suite. If you start from scratch, you may be able to work around the peculiarities of the geckodriver from the beginning. And that is what most people working on other projects tell me, too.
Don’t get me wrong, it it an amazing work people did and do for Selenium. But since Selenium is the current world-wide standard for browser automation it is hard to understand these twists and stunts in the Selenium community. Why wasn’t it possible to first wait for a stable geckodriver before giving up backward compatibility? And even though I personally do not depend on Selenium IDE – many people do. So also here a bit more careful approach would have been preferable. It appears a bit the the Selenium community is not fully aware of the importance their software has for very many companies, and the pain which is created when things stop working suddenly.
Comment by lfrantzen — August 17, 2017 @ 8:01 am GMT+0000
There’s a lot to unpack there, and not much of it is IDE focused. The long and short of it: Selenium is an Open Source project, run by volunteers. Did we want IDE to stop working? No. Were we unaware that lots of people use it? Of course not. Was there anyone — anyone — in the entire community with the energy, time, and commitment to make it continue working? No.
So, come and help. We’d welcome you.
Geckodriver supports the w3c webdriver spec. That’s taken longer than anyone expected to ship. Until the spec becomes stable, there’s no way to implement particular pieces. If you’d like to see improvement, come and contribute to the spec. We need conformance tests, and there are a pile of open issues to resolve. Someone who knows how things should work would help make life a lot easier.
Comment by shs96c — August 17, 2017 @ 9:56 am GMT+0000
Is 55 the actual version when IDE will stop working? The article you link to states “By the end of 2017, and with the release of Firefox 57, we’ll move to WebExtensions exclusively, and will stop loading any other extension types on desktop. ” I’m not sure if there is something else in 55 that will cause IDE to stop working. Thanks
Comment by Rob Mason — August 14, 2017 @ 7:32 am GMT+0000 |
There are other changes landing in Firefox that also prevent IDE from working. 55 is the version in which IDE stops working.
Comment by shs96c — August 17, 2017 @ 9:56 am GMT+0000 |
Yes I agree with the “shs96c”. Geckodriver is more widely used than the other drivers..
“The end of Selenium IDE ” – Naaa… this is just because of change of new api “Web Extension” mechanism (Change is always good..)
Comment by Ganesh Mandala — August 14, 2017 @ 9:32 am GMT+0000 |
It’s very sad news for us that IDE will no longer exist for Mozilla Firefox 55. 😦 😦
I have feeling that in next 10 years Selenium will no longer exist for Firefox. People from Mozilla should come forth and take some steps or give some boosts to Selenium enthusiast.
I would love to contribute at IDE.
I want to join the developers community.
Comment by Avinash Mishra — September 21, 2017 @ 9:39 am GMT+0000 |