View previous topic :: View next topic
|
Author |
Message |
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Hi,
Could anybody tell me approach for regression testing , describing with any example,with any tool?
Expecting a reply soon ! |
|
Back to top |
|
|
rajakishore
New User
Joined: 04 Apr 2005 Posts: 10 Location: Hyderabad
|
|
|
|
Hi Rachel,
Regresion Testing
After u make code changes(not the logical part) to make code base content
the output generated(program funtionality ---Output) after code changes n before code changes...when r equal.....
It is a quality control measure to ensure that the newly modified code still complies with its specified requirements and that unmodified code has not been affected by the maintenance activity.
Example: Dead(non executed) code removal
Regards,
RajKishore. |
|
Back to top |
|
|
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Not able to understand !
After u make code changes(not the logical part) to make code base content
the output generated(program funtionality ---Output) after code changes n before code changes...when r equal..... |
|
Back to top |
|
|
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Hi,
Waiting for a detailed explanation by somebody.
Moderators please help!
|
|
Back to top |
|
|
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Hi,
Why isnt anybody responding to my query ?
Please help me with the answers.
|
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
First you need test data that exercises all rountines in the program being changed.
Next you execute the program with the test data and save all output (new files, modified files, print and screen).
Modify program, and run test data again saving all output from this run.
Compare output of both runs. All unmodified areas of records, reports, screens should match exactly. Modified or new records, reports, screens should produce results that are verified as correct.
Sounds straight forward and easy, but it's best to have all documented in a test plan identifying input, procedures to follow and how and by whom the output is to be verified. Usually, involves redirecting output to sequential files and using a file comparison tool (Comparex, File-Aid).
Finally someone responsible signs off and program is returned to production.
Can be daunting when modifying many programs or an entire system(s).
I've been through VSE to MVS conversions as well as Y2K projects and a plan, the right tools, testing team (users and techies) and signoffs are the only way to go.
Dave |
|
Back to top |
|
|
rajakishore
New User
Joined: 04 Apr 2005 Posts: 10 Location: Hyderabad
|
|
|
|
Hi Rachel
Sorry for replying u late .....
when u go for regression testing is after u make some code changes....
This code changes can be either change in coding standards put forward by the client or can be some Dead Code removal.....
After the above changes are made to the code(its called Convereted) u test them using the same input file which is used before any such changes have happened to the code(its called Baseline) so the output generated by such converted run and baseline run should be equal, we do that by making use of compare and the return code = 1 that means there are no changes in the both the outputs......
So adding to my last mails last sentence ...."It is a quality control .....by the maintenance activity"
I guess u will get a idea of what i mean by this..... |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
Rachel,
why are you not acknowledging responses when you were so impatient to begin with?
Dave |
|
Back to top |
|
|
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Hi,
Thank u very much for the answers and sorry for the delay in responding and i was not very impatient in getting answers .
I m still waiting for more responses for the question i have asked .
Thank you !
|
|
Back to top |
|
|
che
New User
Joined: 31 Mar 2005 Posts: 6 Location: Chennai
|
|
|
|
Lots of tools are used on regression testing. Some of them are as follows,
Hyperstation plus, Hyperstation, Mercury Qtp etc are some tools.
while Qtp uses a record,edit,play mechanism to test the mechanism,Hyperstation and Hyperstation plus use scripts which are written manually to play through the screens for a compare.
General mechanism for Mercury QTP is to load the test conditions and test data on to a database and supply for test runs.
Lemme know if any more info is needed on the same.
|
|
Back to top |
|
|
rachel
New User
Joined: 09 Mar 2005 Posts: 13
|
|
|
|
Hi,
Thanks for the names of the tool !
Can anybody help me out with the material Hyperstation plus,
Hyperstation, Mercury Qtp etc.
|
|
Back to top |
|
|
|