SPONSORS:






User Tag List

Results 1 to 5 of 5
  1. #1
    Member
    Join Date
    Feb 2000
    Location
    san jose,ca,usa
    Posts
    69
    Post Thanks / Like
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    SilkTest\'s timing problem

    I am testing the page rendering time by a browser.

    Here is how I do this test:
    First, let the browser has the cache on and view the page
    to be tested.
    Now turn off the cache. And push the back, forward buttons
    on the browser. The browser should read the page from
    its cache and display it on screen.

    I used SilkTest to push the buttons (repeat 10 times)and find out the time to do it. When it is done, calculated the
    average time per page which should be the page rendering time.
    The problem I have now is SilkTest's result is quite different
    than when I used a stopwatch and munually click the buttons.

    SilkTest's time is faster when I tested a complex page.
    SilkTest's time is slower when I tested a simple page.

    I noticed SilkTest waits quite a while (1-2 seconds) before
    it clicks the next button after the page is fully displayed.
    So I thought SilkTest is too slow. But when I did it
    manually myself. I clicked the page after the page is done
    and it turns out I am quite a bit slower than SilkTest
    ( this is on the complex page case.) I am puzzled.



  2. #2
    Senior Member
    Join Date
    Feb 2000
    Posts
    1,497
    Post Thanks / Like
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Re: SilkTest\'s timing problem

    No specific solution here but you'll find that you can control Silk's timing to a considerable degree by experimenting with seriously reducing or eliminating all of the Agent/option delays and retry increments. And by unchecking all of the automatic tests for enabling, exposure, etc.

    Another consideration is that the operating system itself (NT assumed) performs its own caching; an example of which is a File/Find against an entire hard disk - the first can take quite a while, the second is very fast.

    One question: What are you using to determine when the refresh is complete? One issue here is that the .inc files include more timing delays for browsers than were present in the code for native Windows apps. Using some form of handshake is always far preferable to any kind of blind time-delay. IIRC, these can usually be found in debug mode with the Agent calls turned on.

    Another minor point is to remember that Silk consumes resources too which will affect your numbers to some extent.

    That Silk can beat your manual efforts is a good sign. Try starting the Task Manager and watch CPU utilization while the test runs. It's sometimes useful to see who's using up the clock.

    John

  3. #3
    Member
    Join Date
    Feb 2000
    Location
    san jose,ca,usa
    Posts
    69
    Post Thanks / Like
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Re: SilkTest\'s timing problem

    I already reduced the retry increment.
    I should try to uncheck those checkings.

    I don't have any way to determine if the refresh is done.
    SilkTest checks if the app is available, I assume when
    the refresh is done, the app. is available.

    My script just keeps fliping the buttons without any
    timeout in between.
    Which .inc file has the time delay?
    And what kind of handshake can we have ?

    Thanks for the useful information.

    Ted

  4. #4
    Senior Member
    Join Date
    Feb 2000
    Posts
    1,497
    Post Thanks / Like
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Re: SilkTest\'s timing problem

    Sorry. By 'refresh' I actually meant repaint - the point at which the browser has completed putting all objects and graphics on the screen. The lack of consistent interface stardards in HTML and Java have made this determination problematic. You'll have to find what is trustworthy for your particular application.

    You'll have to dig through the browser inc files to find time delays that may affect your operation. Hence the comment about debug mode for experimentation. Make a backup before you change any original values. This approach is suggested only as a last resort.

    By handshake I mean avoidance of time delays wherever possible in favor of some verifiable event such as the appearance of an object. It also implies logic that will continue to retry a command until the expected event appears or some time limit is exceeded to indicate that it won't. With this approach, you gain the shortest possible overall run time, while ensuring that expected delays are handled in a reasonable manner.

    An example would be a menu selection or a button whose enabled/disabled state is affected by link speed or a tenuous connection to some required device.

    Use of a time delay here becomes a high maintenance item; first to find a workable value; and everytime afterward for any change in the test enviroment that might affect that time value.

    A far better, low-maintenance approach is to create a class for time sensitive objects and imbed retry and timeout logic into their class structure. But I digress somewhat.

    Another thought. CaptureBitmap can be used to determine if a panel is complete as long as flashing or moving graphics aren't involved.

  5. #5
    SQA Knight
    Join Date
    Jul 1999
    Location
    Grand Rapids, MI
    Posts
    3,727
    Post Thanks / Like
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Re: SilkTest\'s timing problem

    You did not mention what browser you are using but I have found that IE 3 and IE 4 browsers interact much quicker with Silk 5.0 then doany of the Netscape 4.x browsers.

 

 

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

vBulletin Optimisation provided by vB Optimise v2.6.0 Beta 4 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
User Alert System provided by Advanced User Tagging v3.0.9 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Questions / Answers Form provided by vBAnswers (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
vBNominatevBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Feedback Buttons provided by Advanced Post Thanks / Like (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Username Changing provided by Username Change (Free) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
BetaSoft Inc.
Digital Point modules: Sphinx-based search
All times are GMT -8. The time now is 05:31 PM.

Copyright BetaSoft Inc.