0

Currently I am using Thucydides & JBehave to perform test cases by writing a feature story file which is pre-defined. Is there any way to perform test on runtime data using storyfile? If not, can you please suggest me a way to do so? Thanks in advance.

Martin Spamer
  • 5,437
  • 28
  • 44
Md Sajid
  • 131
  • 1
  • 1
  • 13
  • Your question is lacking some of the required information. Can you please post a [Minimal, Complete, Verifiable Example](https://stackoverflow.com/help/mcve) or a [Short, Self-Contained, Correct Example](http://sscce.org/)? Thanks! – Will May 11 '15 at 05:14
  • Suppose a user will register for a service. and he will get an email if he registered successfully. Before sending the email to user our local server will calculate data/generate data based on registration form, this data is generating in run time, I want to perform testing on this runtime generated data. (I cant predict them). is it possible to test using storyfile? or with other tools? – Md Sajid May 11 '15 at 05:28
  • To perform two step email validation, use something like Mailhog (https://github.com/mailhog/MailHog) on your test environment to capture the sent email and query that to find out the generated token. As to the general data question, generate the data to a file and load it at runtime http://thucydides.info/docs/serenity-staging/#_using_test_data_from_csv_files – Martin Spamer Aug 26 '16 at 10:21

1 Answers1

0

Actually is pretty much mandatory to gave good random data generator in your Testing framework. In order to avoid the static data and later suffer from it. It's also know as Pesticide Paradox:

If you keep applying the same pesticide, the insects eventually build up resistance and the pesticide no longer works.

This article can explain it additionally.

Another working option, from my experience is the Oracle. I agree that building such complex functionality at first may seems like an overhead. But it works very well with the dynamic data generators.

I'll share some summary after using such approach for 8 months. More then 10 million test cases per run, the actual limitations comes from the computing power that you have, no regression without finding bugs(often in large amounts).

ekostadinov
  • 6,880
  • 3
  • 29
  • 47