I am writing a test that verifies that a warning to the user before the user opens a document. If the user says that they still would like to open the document, the document opens in an external application (pdf or word).
But now I have an external application over the browser window, and it messes up for other tests.
So, what are the best practices around this kind of issue? Rewrite of the appliction to allow for not opening documents in test?
Added description:
The problem is twofold.
1) It starts processes (word and acrobat) that fills the desctop and requires resources from the test-slave
2) The external process seems to interfer with other tests since (guessing here) it is located over the browser window.
what i understood from your post is, the document(word/pdf) is opening in the browser window hence you are not able to proceed with further steps. If so, you can verify the Title to make sure the document is opened in browser window and can navigate back using below snippet.
driver.navigate().back();
Hope this helps.
What I understand from the line
'But now I have an external application over the browser window, and it messes up for other tests.'
is that once user clicks on open button a new window opens up (Window based application) since you have mentioned PDF or Word.
You can use robot class in such cases, below code snippet will close the current active window:
Robot rbt = new Robot();
rbt.keyPress(KeyEvent.VK_ALT);
rbt.keyPress(KeyEvent.VK_F4);
rbt.keyRelease(KeyEvent.VK_ALT);
rbt.keyRelease(KeyEvent.VK_F4);
Make sure you deal with sync issues properly so that intended window is closed instead of AUT.
Related
In my Selenium project word document is being opened as a report.
How can I identify whether the document is opened or not using Selenium webdriver.
In short, unless the document opens utilizing the browser, then you can't with Selenium WebDriver. Selenium only has access to what the browser itself has access to. If the browser is handing this off to the OS or MS Office to handle the specifics then you lose access to it.
You could try to intercept the communication and determine when the code to call the word document to open is executed, but that would be custom code and depend on the application itself.
Personally I think the best way would be to instantiate an OpenQA.Selenium.Interactions.Actions class instance and then specify the specific keystrokes to capture the screen and save it. Then switch back to the browser window driver.switchto().window() and continue. Then you can verify manually that the screen capture has the word document in it after the test is completed.
I have very weird issue with some VBA code. The code is InternetExplorer automation and it's really simple: I just need to load page, enter credentials and click on the button on another page (after successful login).
But... My code is not working (but only on one of my machines on AWS). On my local machine this code works fine.
The weird thing is that it seams I have access only to the Document property of the first loaded page (where I need to enter login/password). I mean from MyBrowser.Document property I can see all INPUT fields when I load start page. But after successful login (the IE window is visible) I see same INPUT fields from MyBrowser.Document property! Also (as I said) I have no issues with this code on another maching.
IE Protected mode is disabled (this is IE 11). I think this is some kind of security issue but I can't locate it by myself...
Here is my code:
'MyBrowser is IE instanse
'Here I'm loading start page and input login/password
'Next the browser show me another page where I need to click a button
'But Debug messages show me input fields for the first Form!
Application.Wait (Now + TimeValue("0:00:05"))
Do
DoEvents
Loop Until MyBrowser.ReadyState = READYSTATE_COMPLETE
Set HTMLDoc = MyBrowser.Document
For Each myHTML_Element In HTMLDoc.getElementsByTagName("INPUT")
If myHTML_Element.Type = "submit" And myHTML_Element.Name = "BUTTONNEWJOBS" Then myHTML_Element.Click: Exit For
Debug.Print myHTML_Element.Name
Next
UPDATE
I don't know why but my browser object is always Busy that's why I get old page HTML. I tried to .Stop it but with no luck.
I don't have answer for IE11 automation, just a hint for a path you may choose to walk.
Another Stack Overflow question Getting source of a page after it's rendered in a templating engine? provides some browser automation options.
One of the mentioned tool has currently broken support for IE11 due to a IE11 behavior change, see http://docs.seleniumhq.org/about/platforms.jsp#browsers. The blocking issue has some configuration hints you may find useful.
In my opinion Internet Explorer was ALWAYS the least reliable tool both from the point of view of rendering compatibility, programmability and even the automation problems accross different versions. And although it had improved over the years Internet Explorer is still better to be avoided (from programmer's point of view). Today it is luckily not the only tool available.
So if you just need to get the work done, there are other browsers or browser-like tools that you can use.
If you MUST use IE then you should get your answer at the Internet Explorer Dev Center → Community → Developer forums
EDIT (after comments)
By the symptoms in your question it looks like the browser object is busy because of some dialog box (perhaps a security prompt) is being show to the user. Some links from Google that may help:
required IE configuration from the selenium browser automation engine - https://code.google.com/p/selenium/wiki/InternetExplorerDriver
how to disable Internet Explorer Enhanced Security Configuration - http://4sysops.com/archives/four-ways-to-disable-internet-explorer-enhanced-security-configuration-ie-esc/
Internet Explorer Security Zones and Compatibility View - https://help.blackboard.com/en-us/Learn/9.1_SP_14/Administrator/030_Browser_Support/030_Browser_Support_IE_Issue
set of dialogs that can be manipulated by the iMacros Internet Explorer automation toolkit - http://wiki.imacros.net/Browser_Automation#Dialog_Manager
Rewriting ~1000 lines of code to use another automation interface (or browser) and asking for help at the Microsoft's Internet Explorer Dev Center are IMO still valid options
Based on your following statement
I see same INPUT fields from MyBrowser.Document property!
it would seem that the page hasn't fully loaded. Perhaps it is loading data via AJAX and clicking the submit button shows the exact same page (which would explain why you were seeing the same INPUT fields) while waiting for a response from the server.
An easy way to test this is to wait for a longer period of time (e.g. 30 seconds) and ignoring the value of the browser ReadyState. However a better way would be looping until you find a element on the successful page that isn't on the first page, with a possible timeout of maybe 30 seconds.
I've used Selenium to do some automation and ran into similar problems and had to resort to using Implicit Waits but I'm not sure if VBA as such a feature (as I don't know VBA)
I'm using the Selenium Client drivers to run tests built in C#. I'm having a problem where the test navigates to a page and clicks on a button in a form with its target set to _blank causing it to open a new window. However the new page being opened returns an XML document and not a typical webpage. Selenium seems to have trouble with this because it hangs when the button is click and the new window opens. No instructions after the click method are executed. The test eventually fails with the error Timed out running command.
Any help/guidance would be appreciated. I've scoured the net but haven't seen anyone running into this particular issue of the page being opened not being a typical webpage which I think is the core of the problem as Selenium can't really manipulate this new opened window. I would post code except that literally all I'm doing is calling the click method on a button which causes a new window to open. Thanks in advance.
To tackle the issue of opening in new window. I typically get url of the link which is to be opened and instead of using selenium.click(), I use:
selenium.open("http://yourwebsite/yourlink")
This helps me to be on the same window only(which avoids opening new window or tab) and when the testing is finished on this page and you want to switch back to original page you can again use:
selenium.open("http://whatever/")
So I'm writing tests for a web page, and one of the tests include clicking a link, which opens a new window, do some assertions in that window, and jump back.
To open links I'm using the technique described here.
The weird thing is, when I go through the steps manually (just press all of them one after another) it works fine. But when I try to run the whole test, it either does nothing (no windows open) or it clicks another button (and not the link).
What can be the reason for it working when i do the steps manually, and not when Selenium IDE runs them for me?
To open link, you can use
click | link=link name
selectWindow | windowname
And then perform operation on that window
I have code that opens a new window but I want to be able to edit the same one.
System.Diagnostics.Process.Start("iexplore.exe", "http://www.live.com")
I'm not sure exactly but a good pointer to start off might be to get the handle of the window you're interested in:
http://www.pocketpcdn.com/articles/dotnetcf_hwnd.html
And then separately investigate what interop messages you can send to IE to change the URL in tab X
In order of increasing difficulty and increased control/power:
Send input text to your IE process. Alt-D to focus on the navigation bar, then the URL, then ENTER.
Use MSAA to find the navigation bar and send it text, as above.
Use MSAA to get IHTMLDocument access to the browser, and then programmatically drive the browser with that, and the related interfaces.
I don't know your exact scenario, but if you can host your own instance of MSHTML, or a WebBrowser control, it will make it a lot easier to get the interfaces and do the manipulations mentioned in #3 above; doing that stuff cross-process is fraught with peril.
I just did a web search and turned up a WatiN tool that apparently wraps a lot of this work; perhaps it would be useful for you.
If you are using 2008 there is a feature where you could create a second form and then add a Webbrowser control
the page could then be called by
myForm.show
The page could then be changed with the
Webbrowser1.Url = New Uri("http://www.google.com")
Use the following code:
System.Diagnostics.Process.Start("http://www.live.com")
That is: do not invoke iexplore.exe directly – just let the system figure out which default browser to open.
This may yield two behaviours:
Either it opens a new tab in an existing Internet Explorer window,
or it creates a new window.
The important point is that this depends on a preference that can be controlled within the Internet Explorer application. If a new window opens, then this is the setting chosen by the user – do not try to override it: overriding the user’s preferences is considered bad manners.
If the users don’t want a new window opened, they can simply change that in their Internet Explorer preferences.