Skip to main content

Testing Midlets in MVM Environment

5 replies [Last post]
murali_reddy219
Offline
Joined: 2007-10-25

Hi Vladimir,

How are you? After long time i got a doubt.

Does the meframework supports test cases running( getting executed ) in MVM Environment?

Thanks,
Murali

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
vsizikov
Offline
Joined: 2004-11-16

Hi Murali,

Sorry for the delay with the answer!

> How are you?

Great, thanks!

> Does the meframework supports test cases running(
> getting executed ) in MVM Environment?

I'm not totally sure what do you mean by "MVM Environment"?
Could you describe it a bit more?

Thanks,
--Vladimir

murali_reddy219
Offline
Joined: 2007-10-25

Hi Vladimir,

MVM Environment means Multitasking Virtual Machine Environment.

I have a use case in MVM scenario, I can download two test MIDlets, execute and post the result which may be having same application ID.If one MIDlet gives result as fail and other one gives result as pass,total result for that test case would be pass(according to MeFramework architecture).

But actually I think we need the result in that case must be fail (when consider with MVM case).Do you agree with that?

plz give me some idea in the above case.

Thanks,
Murali

vsizikov
Offline
Joined: 2004-11-16

Hi Murali,

> I have a use case in MVM scenario, I can download two
> test MIDlets, execute and post the result which may
> be having same application ID.

Hmmm, I'm not very well familiar with MVM environments, and and maybe
I"m missing something, so please correct me if I'm wrong.

For ME Framework, it is an implicit assumption that on a single
device there will be only run test executing at any moment.

There are good reasons for that. This way, we minimize
test interference, which is critical for conformance testing.
Also, some resources just can't be shared (sound sybsystem
might require exclusive access, networking sybsystem might
allow only one or two open connection, etc).

What we typically support is running tests on parallel
devices. Each agent on each device gets it's own ID to distinguish
between them.

It's been done by appending the some unique ID at the end of URL:
http://host:post/test/getNextApp.jad/ID=one
http://host:post/test/getNextApp.jad/ID=two

or something like that.

Would that be applicable to MVM environment?

How do you start test execution in such environment?

Thanks,
--Vladimir

murali_reddy219
Offline
Joined: 2007-10-25

Hi Vladimir

Thanks for your reply .As MVM , we would be running more that one midlet on single Virtual Machine Instance .We are not considering the use case where midlets are executed on different devices.
My use Case is
TCK Midlet1 is downloaded and excecuted , i want Midlet1 to again downloaded and excecuted. I mean we are running two Autotester so scenario may be

Autotester1 running Midlet1 (Test ID 1) May give Fail or Pass

Autotester2 running Midlet1 (Test ID 1) May give Fail or Pass

Both Autotester1 and Autotest2 will be running on same device , on same VirtualMachine .

if Autotester1 running Midlet1(ID=1) gives pass and Autotester2 running midlet1 (ID=1) gives pass result then overall test result at javaharness should be pass , if any one midlet fails then the overrall test result at java harness should be failed. Logically it is AND between the testcases results ..

I think the meframework need to cache the test result from individual autotester (i.e TCP Connection) and update the harness accordingly

Hope i am clear ..Thanks again ..

Regards
murali .

Thanks

vsizikov
Offline
Joined: 2004-11-16

Hi Murali,

Can you specify different URLs for two different autotesters executing on the device?

If you can, than specifying different IDs at the end of the URLs would allow to
execute tests with two autotesters running simultaneously, but without interfering
with each other test results.

From Test Harness point of view, once it gets the test result report back, the test
is considered finished, and it doesn't expect any additional info about it (from different
autotest, or from different device). So, every test is executed exactly once during
every test run.

May I ask: What's the main reason for you to try to use multiple autotesters on a
device? Performance?

Thanks,
--Vladimir