ContributionsMost RecentMost LikesSolutionsRe: Does anyone know what the error message "The tcCrHost.exe process crashed" means and how to fix? Does anyone know if this has been fixed in the latest release (or the one before)? I also experience this a couple of times a week and consider upgrading to a newer version (for this reason among others)if it is fixed. I mostly see it in the beginning of the test when starting up the browser, especially when running virtual mobile browsers (this could also be due to the fact that this involves starting Chrome right after it was closed). /Søren Harder, QA/Test, Zmags Re: Who does your Automated Testing? QA or Development? In my company: Developers do the unit test and QA do the automated functional regression tests in TestComplete. (And developers are responsible for the manual functional non-regression tests together with code-review). You need some development knowledge to do the functional tests (I am far from a developer per se), but I also think you need to have some specific skills for this specific task, which developers will spend as much time to acquire as I have spent. In my company, I do not see that task can be split into a 'coding' part and a 'non-coding' part, where the 'non-coder' defines tests (e.g. in Excel) and the 'coder' writes the basic building blocks of these tests: it will just be running the same tests again and again, but with slightly different values. So based on my experience: have a dedicated person who is has at least basic coding capabilities, but is a QA more than a developer. Re: Running TC scripts on multiple browsers without locally installing browsers. This is through IEs user interface. I have a function without arguments 'IE_switch' that keeps track of a ProjectSuite variable registring how many times IE has run and does nothing first time, runs IE10 emulation second, IE9 emulation third. This uses a function with emulation settings as parameters, which do the work through IEs user interface: opens the dev tools and then the emulation tab, drags it larger, interacts with the drop downs and resizes it to small again. Some of this work can be done with keyboard shortcuts. I also in some tests use this second function 'directly' to test against IE8. Re: Running TC scripts on multiple browsers without locally installing browsers. One way of handling multiple versions of IE is to use the built-in emulator in IE. I have a small function that ensures that if IE is called repeatedly in a browser loop, it goes through IE10, IE9, IE8 in the emulator. I control the emulator through the UI as I would manually (F12, Ctrl-8, click menus, drag size etc). I am aware that some people will not like this, as it relies on emulation and does stuff in the UI that should be handled at lower levels, but I have relatively good experience with it: it has caught a lot of bugs, which were real bugs, and has not let any through. It is not 100% robust though. TC12 slow Property Checkpoints I am currently upgrading my tests from TC11 to TC12 and notice a debilitating change in the behaviour of Property Checkpoints. Tests are getting slower (one test that runs in 7:13 mins in TC11 takes 17:30 mins in TC12). The delay is due to (some) Property Checkpoints that take (exactly?!) 10 secs longer to run. I've managed to remove the problem one place in my test. (Inspired by this workaround https://community.smartbear.com/t5/General-Discussions/Property-Checkpoint-taking-15-seconds-to-verify/td-p/93565 . This is from 2014, which indicates it relates to TC9 or 10, but I did not have this problem TC11). I checked for a property, which was in the DOM, but on a hidden drop-down. By hovering the control to make it visible, the time to do the Property Checkpoint went down from 11-12 secs to 1-2 secs. A similar but more serious problem occurs when I have a Property Checkpoint that waits for an element. I have checkpoints that usually take 2-5 secs, most of which was actual waiting time for the condition to be the case, which now takes the same amount + 10 secs. I even have a test that times animations that take 4 secs, which is hard when the check takes over 10 secs :-s Most of my checkpoints uses the ClassName as property to check for. I tried to change my auto-wait timeout (in every possible way) as I noticed that it too was set to 10 secs, but this did not fix the problem. Has anyone else seen this performance bug? Is there a work-around or a patch? Yours sincerely, Søren Harder, QA/Test engineer (automation), Zmags Re: TestComplete with Jenkins 1) It tells you, you are using all your licenses. This may be because you only have one and it is already running: Jenkins starts TestComplete/TestExecute and cannot do that if it is already running. So turn off TC/TE before starting it from Jenkins. 2) I am a bit in doubt about the other warning: try to use the 'Run interactive user session'. But it may be that there is a catch 22 here, that the Jenkins/TestComplete integration only works when it has its own session and so you cannot run it on the same machine as you use as your Jenkins server. You may want to ask SmartBear-support about this. /Søren Harder Re: TestComplete with Jenkins A quick answer (my weekend has already started:smileyhappy:): 1) Project should not be "GUITestingAutomation\GUITestingAutomation.mds" but just "GUITestingAutomation" 2) I believe you are getting the error code -10 because of the '\LogonAndExecute" which you get because you set "Run Interactive user session" (I understand you have tried without it; which error do you get then?) /Søren Harder, QA/Test engineer, Zmags Re: TestComplete with Jenkins I can see that the exe-file is in 'TestComplete 12/bin' (in the output when you are using the integration), but in the batch-file you call it in 'TestComplete 11/Bin' (bin vs. Bin and 12 vs 11). Re: TestComplete with Jenkins Concentrate on the first line, the rest is just telling you it cannot read the output of the process it cannot run. Try to disable the Run Interactive user session. I believe it is trying to create a user session on the machine where you already have a user session. If that does not work it is possible that the Jenkins integration does not allow you to run everything on one machine. Then I would probably try the batch approach you had in your first mail. I will take a look at it again to see if I notice something. Re: TestComplete with Jenkins The output you get, lists your installations: You seem to have a TestComplete, but no TestExecute. Try to set it to run on TestComplete instead of TestExecute.