[omniORB] OmniORB2 test harness?

[Sai-Lai Lo]

>>>>> Harri Pasanen writes:

> Does the exist a stress test, or any other tests for omniORB2 except the
> demos?

I've uploaded most of the omniORB testsuite to the COST web site.  However
the source is targeted for omniORB3, i.e. it uses POA.  The site is still
at its temporary home and I'm told not to disclose the address publicly
until it moves to its permanent home. In the mean time, if you or anyone
wants to try it out, please drop me an email.

For your purpose, I think the most useful bit of the test is the collection
of IDL samples. Frequently, we find that we have to tweak the 
stub code generation to cope with different bugs and deficenies in the C++
compilers. It would be nice to test KAI C++ against the collection to see
if there is any hiccup.

> I've ported omniORB2 to KAI C++, and would like to compare the
> performance/reliability with that of other versions.
> Maybe just something that excercises the naming service heavily.

The testsuite has one test which I find useful to verify if a compiler
is doing fine with multithreaded exception handling. It is so simple that
you might as well write it yourself. Just take an echo example client,
change the code so that multiple threads concurrently pounding a
server. Then kills the server while the calls are in progress, the client
threads should see a COMM_FAILURE exception and cleanup properly. This is a
tricky test for the compiler's exception handling because all the client
threads throw COMM_FAILURE almost at the same time.

Regards,

Sai-Lai 


-- 
Sai-Lai Lo                                   S.Lo@uk.research.att.com
AT&T Laboratories Cambridge           WWW:   http://www.uk.research.att.com 
24a Trumpington Street                Tel:   +44 1223 343000
Cambridge CB2 1QA                     Fax:   +44 1223 313542
ENGLAND