I wrote a function that gets called for every line a large xml file. When I run in the normal development environment
run mode verify, the test takes 5 minutes 29 seconds for a particular xml file.
Same test same xml file, when I run minimized takes 1 second. This is even more confusing, when I take this empirical data (which I was very fortunate to stumble on) and then look in the WR User Guide (and the TSL guide which frankly this should not be in).
Keep in mind I am more upset at how to get this kind of stuff straight? Is there a referance on this that I missed?
Does this issue even make sence to anyone .. I can see in the normal development environment (non minimized) the flow from main to function and back again slowing the whole show down.
At the very least perhaps we could put this in the Posts gotcha guide or FAQ or something like that (my apologies if we already have). I must say it's quite exciting to go back to Winrunner after several years, it's also a tad frustrating .. but never dull.
It is due to the animation of the script. When you run the script in normal mode, each time a line is executed, you can see the yellow arrow move to the next line until you reach the end of the script. When you are in minimized mode, this action doesn't occur, and speeds up test runs emmensely (as you have noted already).
The only reference I know off hand directly to this (not on my work system, so no documentation), is in the User Guide's chapter related to running from the command line (Chapter 27?). There is an option -dont_animate you can call which means it will open full page, but not update the arrow position while it runs.
Just one of the side effects of a non-compiled language and thrid party IDE that has been modified to fit the needs of WinRunner.
It shouldn't impact your test runs with timings, as you normally (or atleast in all the cases that I am aware of and have been involved in), run with the tests minimized. This leads to a real need for testing of synching with the app by performing the test in minized mode prior to releasing the test to real "production" for your regression suite.
These are just things that you learn when working with the tool, and since no new real development is going to be taking place with WinRunner, I doubt youa re going to see any leaps and bounds in the documentation of what is out there for it now.
Tony as always my thanks! I really would have been mad at myself if you said .. hey dude it's in chapter 3 under ways to run scripts, and consequences. Frankly this is MI'
s bad! I really wish MI and Segue for that matter would sponsor groups to write a companion doc for thier products!
I mean really what a wealth of knowlege these corporations IGNORE! I also agree that for the most part a function that is called a few times in a test script wi animation is NOT gonna clean the clock.
But for those times when you have a function called hundreds or even thousands of times .. preety BAD!!!
(again 3/5 exclamation points .. bordering dangerously on 4)