Forum Discussion

clay's avatar
clay
Contributor
8 years ago

Script conversion takes several minutes.

As a general practice, we would like to convert all of our keyword tests to scripts.  Most conversions just take a few seconds at most.  However, some take several minutes to compile.  The resulting ...
  • clay's avatar
    clay
    8 years ago

    I decided to remove all of the visualization images from my keyword test.  Now the script conversion is almost instantaneous rather the 10-15 minutes.  I do not understand why the image snapshots would have such an effect on the script conversion, but the impact is apparently huge.  Perhaps TestComplete developers could fix this??

     

    Thanks to all for the help.

cunderw's avatar
cunderw
Community Hero

Can you give a little more info? What at the differences in your KWTs? Are there a lot more parameters? Visualizer images? 

clay's avatar
clay
Contributor
8 years ago

We don't use parameters very often.  Our script building approach is to first map out the procedure using the recorder with the visualizer turned on.  Then we step-by-step refine the recorded steps, adjusting the name maps and trying to use coordinates localized to the objects and etc.  As we refine, the number of images goes down because we are replacing recorded steps.  The last step is to convert to script

 

I guess that means there are a lot of visualizer images, but from key word test to key word test I don't think one has more than another.

  • Marsha_R's avatar
    Marsha_R
    Moderator
    8 years ago

    Do the longer ones happen after you've been working on them for a while or randomly during the day?  If they're always towards the end, I would suspect a TC memory leak.  

    • clay's avatar
      clay
      Contributor
      8 years ago

      This is not a problem that gets worse.  The first time and every time I try to convert certain scripts it takes several minutes to compile.

      • tristaanogre's avatar
        tristaanogre
        Esteemed Contributor
        8 years ago

        Without more detail about what is in the KWT that take a few seconds and what is in the ones that take minutes, not much we can do.  We'd probably need to see the actual KWT and everything that goes along with them to make the best evaluation.

         

        It's possible someone from SmartBear would know more about the processing that goes into that conversion and what could cause differences.  But they, too, would probably need to know more about the test cases you're working with to give you a clear answer. 

         

        My best guess is that there is something that is in the test cases that take a while that needs extra processing when going through the translation to script code.  Test cases that are simply "click, keys, settext" on onscreen objects probably translate a lot easier than conditional statements, while/for/do loops, etc.  The simple actions are a straight call to the syntax engine of the language.  The other code is a bit more complex since each language has it's own way of syntaxing such loops.  

         

        Just a guess, but that's where I'd start looking.