Forum Discussion

Fred_at_Remsoft's avatar
Fred_at_Remsoft
New Contributor
9 years ago

How do I do a region compare with different resolutions between record and replay?

I'm testing a desktop application, developing the tests (using testcomplete) on my PC, and then executing them (with testexecute) on a series of remote machines.  I need to do a region compare on a m...
  • Ryan_Moran's avatar
    Ryan_Moran
    9 years ago

    You are exactly right that the different sized window will be too different to compare, but you could force the window to a specific size, like 800x706 (to be small enough to fit both screens) and just compare within the window.

Ryan_Moran's avatar
Ryan_Moran
Valued Contributor
9 years ago

Seems like you should be able to set the resolution on your VM so I would just suggest investigating that further.

If the resolution is different between the two machines then your image will not be the same, and thus region compare will show differences. Alternatively, you can run and update the region compare on the VM and NOT update it on your computer with the different resolution. Ultimately though it is as you said that two different resolutions will produce different images.

  • Fred_at_Remsoft's avatar
    Fred_at_Remsoft
    New Contributor
    9 years ago

    Thanks - I did my best to get the VM to change its resolution, but the software we're using seems to have one set resolution for some reason.  I have a call into our vendor on that side for that - because their help system isn't NEARLY as robust as the setup here, sadly.

     

    I ended up cannibalizing an external video adaptor from another machine in the office, so I could add a 3rd screen that WOULD allow me to set it to the resolution of the VM exactly, and after some playing around with tolerances, I got it to accept the updated image.

     

    And while that’s solved it for this particular instance - Isn’t there a way to take/capture the results from the failed test, and say “Yeah, that’s close enough, use that as your basis for comparison going forward.” …?

    • Ryan_Moran's avatar
      Ryan_Moran
      Valued Contributor
      9 years ago

      Only two things that I know you can do here:

      You can use the tranparent paremeter and specify an area of the image you want to ignore.

      You can also use the tolerance setting.

      Check out this article on Region Compare.

      • Fred_at_Remsoft's avatar
        Fred_at_Remsoft
        New Contributor
        9 years ago

        Hmm.. I looked at both of those options, but I didn't think either of those would help, as the basic image is a different size (because of the resolution).  On my host PC, it's say, 850x750 ... but on the smaller resolution of the remote machine, it's only 800x706.  No amount of masking, or tolerances (unless the tolerances are so large that it effectively could be a picture of micky mouse on the screen and still pass) will pass the test… right?  Even the masked areas would be different between the two, wouldn’t it?  Or have I misunderstood how the masking process works…?  Is the comparison not relative to the position on the screen, or to the window (which in both cases, would be different x,y coordinates)?  Even if the image is exactly the same – if you offset it by say 50 pixels either way, unless your tolerances are large enough to account for that 50 pixel drift (in which case, really, so long as there’s SOMETHING on the screen it would pass, right?)

         

        Apologies by the way, for the newbie questions – I’m still quite new to the software, and it’s a LOT more robust and varied, as compared to the previous automation software I had been using.  I just want to make sure I’m leveraging the extra capacities here to their fullest.