1

What I'm doing

I'm new to Selenium and trying it out because my boss might have projects I can work on if I get proficient at using Selenium to test our company web aplications. I did some research to identify the main problems that afflict webtesters using Selenium and one of them is reliably finding locators for webelements easily. To rememdy this and make my life easier I've set up a simple program to take in locators (xpath,className, ect) and try each one of them iteratively.

Whats the problem then?

Now that I've got a good solution for trying each possible locator once I have them, I need a quick (preferably less manual) way to obtain these locators. Up till now I've been using the f12 chrome developer tool to inspect the webelements and then manually copy paste the relevant information into the method:

Webdriver driver = new Webdriver; driver.findelement(By.id("id was pasted here"));

However finding the ids and things manually and having to read all that HTML stuff is hard and quite confusing since I haven't really delved into web design all that much so Javascript CSS XML and HTML are relatively new to me. I know some of the (very) basics but I couldn't find my way around a well designed website.

What I want

is to be able to click something like "inspect element" and be greeted with a list of locators for that webelement. Usually the easiest way for me is to rightclick on the highlighted HTMl for that element and select "copy Xpath" but xpath expressions that are auto generated aren't always the best for locating elements. I found a similar question that deals with this issue here:

Selenium 2 : finding web element locators programatically

However that question doesn't fit my needs because it's more specific into flexible xpath locators and doing everything purely programatically. Also the answer provided was "use the developer tools" and "no there are no other tools". I'm willing to try my hand at making a simple tool but I need to know where to start first. What I desire is a simple way to click or otherwise select/ programmatically log a webelement's locators and/or get a simple message box that finds me possible ids, locators, classnames, possible xpaths in a list ect. I want to do this so I can just copy paste the text from that message box or log message into the code that needs to identify the webelement.

Good idea but where do I start?

My problem is I have no idea where to start with this. I think I would have to learn of some way to program to my mouse and then learn HTML enough so I could parse out the locators but I need guidance because that is way to broad and is likely not actually what I need. So hence I ask, what do I need to learn/do in order to be able to see all the possible locators of a webElement all at once easily. I am open to suggestions about how I might learn/create a simple tool that would help me (and probably a lot of other people) overcome this problem.

Improve this question
4
  • You can use Selenium IDE but it works only with FF. It allows you to record, edit, and debug tests and you can configure it to use different languages (python, java, c#) and it give you results in multiple locators (xpath, css) Commented Sep 11, 2018 at 16:17
  • 1
    Keep learning. Don't give up. Unfortunately, this question is off-topic for SO, as it is way too broad. Commented Sep 11, 2018 at 18:33
  • Yeah I had feared that a little. I don't know where it would really be on topic though, its like a really broad programming question. I'd be willing to make it more specific if I knew how but I don't have a specific problem other than grabbing locators easier Commented Sep 11, 2018 at 18:36
  • Maybe I'll change the question to something like "programatically grab locators for webelements" and instead of ask for something this broad ask for ideas on a program to take in HTML and return the relevant information. Would that still be too broad? Commented Sep 11, 2018 at 18:40

1 Answer 1

Reset to default
1

To a degree, the only way you'll get your hands on quality locators is to hand craft them. There are plenty of css selector generator tools out there in the form of browser extensions that let you click elements and get an autogenerated locator in the form of css or xpath. Big but though, you end up getting very verbose and brittle selectors through methods like this.

Those tools are definitely useful, however, because you can take what it gives you as a baseline and then start whittling away until it becomes the level of verbosity you need, while relying on as few unrelated elements as possible. In a best case, you can use firefox's or chrome's element inspector tool to fast-travel to an element in the DOM, and if your element uses a unique attribute like id, then that's often all you need. (You can easily test your locators with the javascript console if the app uses jquery by doing one of $("locator") $$("locator") $j("locator") depending on your jquery version)

While figuring out your locators never stops being annoying, it becomes much easier and in my opinion will boost your productivity in the long run because locators that work now can and will occasionally stop working as the app under test undergoes dev revisions, and finding out what about it is different can be a debugging challenge in it's own. So part of minimizing the occurrence of your locators breaking due to changes in the app under test is getting comfortable manually inspecting the dom and creating locators that are as locationally independent as possible. And what I mean by locationally independent is by targeting attributes that won't change, and by not relying on element hierarchy or sibling element's ordinal value (in the case of list items) which can change as a UI gets revised.

Every element you rely on to locate your target is a liability and contributes to the brittleness of your overall locator, it's impossible to avoid this completely though. That human touch is what lets you be as verbose as you need, while being as concise as possible to have a robust and functional locator

Improve this answer
Sign up to request clarification or add additional context in comments.

5 Comments

While this provides more of a "no, don't bother" solution to my question, I still think this is a good answer. I just think it would be convienient to get ids without having to dig for them, or being able to get a list of possible xpaths and cssselectors so that in the worst case, at least you have a few options (and then maybe you can use a greedy algorithm to decide the best approach).
I just thought of something. What if a tool was developed using the WayBack machine (or similar) to provide a sort of linked revision history (I have some xpath at time t and I know I'm on website W thus when W updates after time t+1 I can use the revision history to see the difference and programmatically update my xpath).
I think it's a really cool idea, but with my past experience I wouldn't expect that to be an efficient way because you will likely still have to go in and make modifications. The more programmatic you go, the more frequently you'll need to make the modifications. A change to the DOM can be in a wildly different place than your target, and at first and second glance, can appear completely unrelated, and sometimes making a robust locator involves targetting something very specific, or avoiding something very specific that a computer cannot understand the notion of when auto-generating.
If your goal is to be effective, your idea sounds more like a job by itself to develop that's separate from using it to develop something else
@Redacted Did you find the answer to the question you raised ? I have one idea but need to discuss with you. Can talk over a call !

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.