Most of the time i have encountered situations where the developer says that the bug is not reproducible in his/her system even if i reproduce that bug in my system (i.e. the QC environment. In such a situation myself being into QC, how am i suppose to deal with this situation.
Ask the project manager to enhance the system requirements for your product: All development tools that the developer has installed on his system should be system requirements to guarantee your product will be working properly... [img]images/icons/wink.gif[/img]
Seriously: Tell the developer that it is not a criteria that it is working on his machine. If your testing environment is a kind of environment the product should be working on, IT IS A BUG.
If this will not lead to a solution, let the project manager decide how to deal with the bug (give responsibility to the project manager).
What we do is set up our test environment on a VMWare image. When a test fails, VMWare stops and takes a snapshot of the image, we then send the snapshot to development and they run it. It will contain all data and operations that have occurred up to the point of failure. This beats trying to prove it failed on the test box.
It's not expensive either $300 per year for the license and the player is free.
Rich - like the approach must save a lot of time in reproducing defects.
To supplement Bru's comment - the applications we test are supposed to work on all machine's - it is doubtful that the end users will use the developers machine and image. Developers have a tendency to forget that they have software and functionality that the rest of do not. I try to nicely remind them of that - fortunately right now this works for me most of the time.
When there are outstanding defects that cannot be reproduced I also work with the client in UAT and ensure that they include these tests - if they pass on their machine it remains unreproducible - if it fails it is now reproducible and gets more attention if the client considers it a problem.
Who cares about a developers env or machine ?
No point trying to reproduce it on their machine for demonstration, get them to see it on the Controlled Test Machines..... BTW if you are not controlling your test environment.... then you have a real problem. Once they see the problem and have done some debugging on the Controlled Test machine..... then they can fix it where ever.
Alternately, there could be something wrong with the build on the test machine that did not happen on the developers (error during install, missing file, bad configuration...)..... whichever way.
I would like to thank everyone for the enlightenment.
Could you please elaborate a little on how to setup test environment on a VMWare image. I wud like to setup the same, here from scratch. I searched in the net and found a link to download a free copy of VMWARE PLAYER.