You would need a separate machine for each user. Unless your application
can have multiple instances of itself running on one machine. Then you
would need to either have multiple GUI maps (1 for each instance) or
write a function that would allow you to reference one GUI map for
You could do a loop, but you are only going to get one user at a time with WinRunner, because it needs to take over the computer to work. So one user per machine at a time. If you put the user in a loop you can get multiple users but they are consecutive not concurrent.
No, they are not based on the same engine. The technologies are
completely different. WR is a gui oriented tool that has no knowledge of
the network connections or internal workings of the application. While
LR has no knowledge of the application gui. LR is only concerned with
the network communications of the app and the server. Actually, once a
"recording" is made with LR of an AUT network transaction (and it's
properly correlated and debugged), the AUT is no longer needed for LR to
do its thing.
User Anne Martin (email@example.com) posted:
I think that he's referring to straight concurrency,
based on what TeamTest used to do, and what you can
still do with Silk, to an extent... Theoretically, you
could run the same scripts with multiple users on
multiple machines using WinRunner to achieve some
level of concurrency - but not a very high one!
--- Richard Weber <firstname.lastname@example.org> wrote:
I believe there is another option to run vmware (virtual machines inside
win2k or NT- www.vmware.com <http://www.vmware.com/> ) but not sure how
the licence agreement will work with mercury.
With typical 256 meg of ram I can create up to 5 virtual windows .
No we don't have LR licenses yet, I was just thinking why it can't be
done or can be done with WinRunner since they based on the same engine.
I was doing something similar before with rational Team test 5 years ago
and was wandering if I can do it with WinRunner