New Idea

Support for Google Polymer framework

Status: New Idea
by Christophe on ‎01-10-2018 01:52 PM



I suggest that TestComplete supports Google Polymer framework.


Thanks in advance.

For now I have to count the number of automated scripts within every folder manually. That's not convenience. If you add the number next to folder, that's good for user. The high-level design is like the enclosed picture.



I would like to have the ability to identify from a script routine what keyword test is currently being run. Currently there is no way.  In javascript I know there is a caller method I could use, but that doesn't get me the name of the keyword test, it gets me the previous methods name.

It would be helpful if the master pc can recover from lost connection.


Currently I have a test which needs to lose the network connection as part of the test and then it establishes the network connection.


Once we lose the network connection the master pc is unable to receive the results from the slave pc.


Currently, we would have to run this test locally on the slave pc as a workaround.


Attached is the screen shot of what happens on the master pc

I don't know if this has been suggested yet or not.  But one of the things that the TestComplete methods indicated in the subject does is that, rather than finding child objects from the first on the page/form to the last (top down in the object browser), they instead work in reverse.


For example, if an application form has 5 buttons with similar information and you use "FindChild" to find one of those buttons, it will find the button that is the last one in the object browser rather than the first one.  FindAllChildren orders the resulting array in a similar "reverse order".


While for the most part this isn't a problem, if you want to find the "first" object on a page that matches a certain set of criteria, these functions present a challenge in that you need to account for this reversal.  I'd like to see the Find methods find objects based upon ordering on screen (top down) rather than the reverse (bottom up).

Currently, the mouse "Drag" method drags from a starting co-ordinate a set X distance and Y distance from that starting co-ordinate.  For most "normal" drag functionalities, this is sufficient.  I even wrote some code with help from folks up here that, given a starting object and a destitination object, will drag from one to the other.


function dragCalculatedDistance(startObject, destinationObject) {
    var startPoint, targetPoint, dragX, dragY;
    // Drag from the center of object A to the center of object B

    startPoint = startObject.WindowToScreen(startObject.Width/2, startObject.Height/2);
    targetPoint = destinationObject.WindowToScreen(destinationObject.Width/2, destinationObject.Height/2);

    dragX = targetPoint.X - startPoint.X;
    dragY = targetPoint.Y - startPoint.Y;

    startObject.Drag(-1, -1, dragX, dragY); 

The problem is this...  We have a situation where we need to drag an object that is present below a table in our application to a particular cell in the table.  In that process, when we cross the lower border of the table, the table may auto-scroll down to reveal other rows.  Because the above function just drags to a calculated set of co-ordinates, the end result is that we end up dropping the object on the wrong row in the table.


Proposed Solution: A drag functionality for TestComplete where, instead of dragging based upon a co-ordinate distance, actually drag from one object to another... something like



where the two objects are identified onscreen objects.  This would resolve the auto-scroll because it would actually drag to the particular cell we want and not to just a set of on-screen co-ordinates.


Low priority change to make... and I'm open to other code that would make this work better for us.  But I think this would be "cool" as it would mean that my custom code above wouldn't be necessary any more and we could have "smart" drag-and-drop functionality.



Frequently, I run into problems where my find()/find_child()/find_child_ex()/etc.. calls are finding the wrong objects (sometimes its hard to find uniqueness within the objects you have at your disposal). It would be AMAZING to have a 'finder' tool directly in the object browser that would allow you to pick an object to perform a 'find()' call on, give a form to fill in props and values and it would give you the object (or objects) that it is matching against without having to run any of your own code.


It's good to know right off the bat if your props and values are mapping to the object you expect. I could see it saving me a whole lot of time when developing test scripts.

I have over 700 test items grouped into groups and distributed across multiple test machines in a network suite, so the result log is long.

I know I can go directly to an error but it doesn't select the test item in the tree (another annoyance).

When I only have a few errors I have to step through the whole tree to find them.


It would be nice if there were a right click menu item in the tree to select the top level and collapse all nodes, or expand all nodes.  Then I could collapse the entire tree, do a couple of high level expansions, and easily see the failing group and test.

Our Windows application has this functionality courtesy of our DevEx tree controls.  Since TestComplete has a similar look and feel, I hope that it would be easy to add this to your existing right mouse button tree menu.



Improvements after renaming/moving Aliases

Status: New Idea
by nebsta on ‎12-07-2017 08:42 PM

Given the technical difficulties of implementing the following idea please can we get an interim solution implemented that will, upon renaming or moving of an Alias within the hierarchy:

  1. Automatically update the Alias of 'Compare Property' parameters when they are set using Mode = Object Property
  2. Automatically update the Alias of stores objects (e.g. DB Table object).
  3. In order to capture other occurrences of the Alias within scripts etc, the user is prompted if they wish to open a find/replace, if they choose to it will automatically populate the former Alias name in Find and new Alias name in Replace with options for the scope to be set to:
    • Current open windows
    • Current Project
    • Project Suite

Script extensions are a very useful feature of TestComplete, allowing end users to create their own custom objects, checkpoints, keyword operations, etc., for use in their automation projects.  We use script extensions in building out large sections of our automation framework, encapsulating functions like ADO SQL queries so that the code is accessible across all projects without needing to include files in every project.


At this time, when running test automation using the Environment Manager interface to CBT, these custom script extensions are not bundled into the project when it is sent to the virtual environment of CBT, meaning that any tests we construct which may use parts of our extensions will be unable to run on CBT.


I would like to see Script Extension support included in the interface to CBT.  Perhaps look at the script extension directories of the TestComplete/TestExecute launch on the local machine, package up any TCX files found in those directories, copy them up to the CBT environment into the TestExecute instance there prior to launching the tests.  This would make using the Environment Manager a more robust solution for testing in multiple browser environments by allowing more customization of code frameworks.

I have used the data generator wizard to create a table of passwords.  It would be nice to be able to re-generate that table on the fly from a script so that fresh passwords exist.

As all supported scripting languages (except VBScript) support optional function parameters that can be set within the function body, I think that it is a good idea to make the 'Run Current Routine' Code Editor's context menu command to be enabled for both procedures and functions for all scripting languages except VBScript (where it should be left disabled for functions as it is now).

Results Display in Team Foundation Server

Status: New Idea
by campk37 on ‎10-26-2017 12:57 PM



So I love that I can run a build through TFS and send out an email with the results attached in a file. However, not everyone that would like to see these results can view and understand the email that is being sent out. Also the quantity of the tests is also in question occasionally. Right now, the data that is recorded by TFS system is very misleading. It makes you think you have only executed one tests. Ideally displaying the results overtime on the Dashboard and accurately reflecting the data within the file that states which tests have passed and which tests have failed would be ideal. 


The functionality that I am imagining is similar to the results display of builds and unit tests that exist in TFS today. 



User Story - 


As a User I want to be able to view test results by test case through the TFS Dashboard System so that I can better display the results of the tests I conduct using Testcomplete. 

Hello All,


I propose to implement table checkpoint for Sencha Grid (Ext JS Grid Panel) control like for JTable objects (Java Swing).

For now we use Regions checkpoints or "if...else" construction for checking Sencha Grid, but table checkpoint is more user-friendly, because it is easier to view differencies during test execution, Regions checkpoint fails often for different computers due to diff column width or something else (I think it's not good itea to set 500-700-so on for Pixel Tolerance).

For Table checkpoint we will get passed tests with correct values in grid regardless of external factors (mouse pointer, diff column width, tooltips, etc.)



Enable property/method which allows checking whether a row is selected in the Sencha (Ext JS Grid Panel) control.

I have a ProjectSuite with multiple projects in which keyword-test-routines of one project are referenced in multiple other projects. If an original keyword-routine is renamed, there is no other way but to remove and then readd the references after which all usages of these references in actual tests have to be found and replaced with the renamed references as well - needless to say, that this is always a quite hideous task.


It would be very helpful if renaming a keyword/script would also rename all references and all places where they are used.



For some projects there is a need to process test logs and export them to separate log files. But if the test was terminated from debugger, such test log must be skipped.

Currently there is no easy and clear way to distinguish between test that was terminated from the one that just failed during runtime.

I would like to ask for some flag or property that will make it clear whether the test was terminated from debugger or failed normally.

We would like to be able to capture any logs from the console or network messages so that we can provide specific network errors to our developers for any intermittent bugs that may occur which are of this nature.

Currently, while we can open developer tools and access this via the object browser to get the elements which contain the log messages, we can't do this solely via an automated run in Test Complete. Is there a way we could get this information and if not could you add a feature for getting it?

Dear all,

I am facing a problem with XML checkpoints. The root-node of my XML file has an attribute in it, which reflects the date, the file was generated. I would need to skip ONLY this attribute during compare/check phase.

I cannot use the "Ignore attributes" option, as this would ignore all attributes on the file. But other attributes are relevant.

I cannot delete the node from the checkpoint (as suggested by the online help about excluding information from the comparison), since it is the root node. All other nodes would be gone as well.


Feature request:
An option in the XML checkpoint editor would be great: just right clicking an attribute, selecting "Ignore on check/compare".

So far, one will first need to alter the actual XML file and manually remove the attribute.


Personally, I hate the new feature in 12.30 "Smarter auto-formatting. When you type a bracket or a quotation mark, the Editor now automatically inserts the corresponding closing bracket or quotation mark."


Please see community post for more details:


I would love to have an option to shut this off.

Suggest Product Idea
We need your help! This is your space to provide suggestions for features or improvements for SmartBear products. Let us know what you think, or how we can help improve your experience with TestComplete.
Top Kudoed Authors