| || |
Hi BetaSoft experts
I am using SilkTest for the last year and currently evaluating performance testing tools.
Part of our offering include uploading files (text, avi, mpeg etc..) and during the upload time, we display to the user information based on the his preference.
I was pleased to see how detailed was the level of recording, but when I showed this to our chief architect he raised some points that I needed some help with.
1. The content of the uploaded file was embedded within the recorded script (I think this is a great feature) but the question I was asked is how can I parameterize the file that needs to be uploaded?
2. If I parameterize the user name and password values, I will login as different user however in my script I can still see the recorded vales which belongs to another user and therefore wrong data is downloaded
Is there a way (patch or API calls) that enable me to bypass these issues
Can you recommend any other workaround
Re: File Upload
Actually you can do both quite easily. Create a variable for username and password, put those in the place where they appear in the recorded script (in the TMain transaction), then retrieve those values from an outside data file during the TInit transaction. Also, you can just submit random username/passwords quite easily if your app will accept them. Look at the sample scripts in c:\program files\segue\performer\working\samples\internet for more detail.
Issue two -- what file to download -- how does your app know what file to serve? Performer can be told in the same manner as your app is told. Also, randomize your selection of files -- make a list of all files, and have your app select any of them according to pre-define probabilities. If uploading is the issue, pull from an outside datafile -- embed the files in an excel sheet. Randomize the order of selecting files, or specify the order before running your script.
Hope this helps,
Re: File Upload
Thanks for your help.
In many cases I need to upload a binary file (2MB) and for this excel is not good.
The second problem is that the script itself includes many images that are based on user_1 but when I parameterize my test to use user_2 a different set
of images should be downloaded but in my script I still see the requests that user_1 made
There is another tool (not sure for name) that records only the requests made to the main pages and not for all the subsequent images